DataAnalysis.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# <font color=#004090>LOAN DEFAULT PREDICTION</font>\n",
    "-----\n",
    "## Machine Learning Engineer Nanodegree<br>Capstone Project Notebook\n",
    "Ricardo Szczerbacki <br>\n",
    "October 2019\n",
    "\n",
    "This is the work notebook where the data analysis was made.\n",
    "\n",
    "\n",
    "\n",
    "### Data Analysis\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>TARGET</th>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <th>CODE_GENDER</th>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <th>...</th>\n",
       "      <th>FLAG_DOCUMENT_18</th>\n",
       "      <th>FLAG_DOCUMENT_19</th>\n",
       "      <th>FLAG_DOCUMENT_20</th>\n",
       "      <th>FLAG_DOCUMENT_21</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_HOUR</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_DAY</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_WEEK</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_MON</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_QRT</th>\n",
       "      <th>AMT_REQ_CREDIT_BUREAU_YEAR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100002</td>\n",
       "      <td>1</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>202500.0</td>\n",
       "      <td>406597.5</td>\n",
       "      <td>24700.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100003</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>N</td>\n",
       "      <td>0</td>\n",
       "      <td>270000.0</td>\n",
       "      <td>1293502.5</td>\n",
       "      <td>35698.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100004</td>\n",
       "      <td>0</td>\n",
       "      <td>Revolving loans</td>\n",
       "      <td>M</td>\n",
       "      <td>Y</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>67500.0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>6750.0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100006</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>F</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>135000.0</td>\n",
       "      <td>312682.5</td>\n",
       "      <td>29686.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100007</td>\n",
       "      <td>0</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>M</td>\n",
       "      <td>N</td>\n",
       "      <td>Y</td>\n",
       "      <td>0</td>\n",
       "      <td>121500.0</td>\n",
       "      <td>513000.0</td>\n",
       "      <td>21865.5</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 122 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  TARGET NAME_CONTRACT_TYPE CODE_GENDER FLAG_OWN_CAR  \\\n",
       "0      100002       1         Cash loans           M            N   \n",
       "1      100003       0         Cash loans           F            N   \n",
       "2      100004       0    Revolving loans           M            Y   \n",
       "3      100006       0         Cash loans           F            N   \n",
       "4      100007       0         Cash loans           M            N   \n",
       "\n",
       "  FLAG_OWN_REALTY  CNT_CHILDREN  AMT_INCOME_TOTAL  AMT_CREDIT  AMT_ANNUITY  \\\n",
       "0               Y             0          202500.0    406597.5      24700.5   \n",
       "1               N             0          270000.0   1293502.5      35698.5   \n",
       "2               Y             0           67500.0    135000.0       6750.0   \n",
       "3               Y             0          135000.0    312682.5      29686.5   \n",
       "4               Y             0          121500.0    513000.0      21865.5   \n",
       "\n",
       "              ...              FLAG_DOCUMENT_18 FLAG_DOCUMENT_19  \\\n",
       "0             ...                             0                0   \n",
       "1             ...                             0                0   \n",
       "2             ...                             0                0   \n",
       "3             ...                             0                0   \n",
       "4             ...                             0                0   \n",
       "\n",
       "  FLAG_DOCUMENT_20 FLAG_DOCUMENT_21 AMT_REQ_CREDIT_BUREAU_HOUR  \\\n",
       "0                0                0                        0.0   \n",
       "1                0                0                        0.0   \n",
       "2                0                0                        0.0   \n",
       "3                0                0                        NaN   \n",
       "4                0                0                        0.0   \n",
       "\n",
       "  AMT_REQ_CREDIT_BUREAU_DAY  AMT_REQ_CREDIT_BUREAU_WEEK  \\\n",
       "0                       0.0                         0.0   \n",
       "1                       0.0                         0.0   \n",
       "2                       0.0                         0.0   \n",
       "3                       NaN                         NaN   \n",
       "4                       0.0                         0.0   \n",
       "\n",
       "   AMT_REQ_CREDIT_BUREAU_MON  AMT_REQ_CREDIT_BUREAU_QRT  \\\n",
       "0                        0.0                        0.0   \n",
       "1                        0.0                        0.0   \n",
       "2                        0.0                        0.0   \n",
       "3                        NaN                        NaN   \n",
       "4                        0.0                        0.0   \n",
       "\n",
       "   AMT_REQ_CREDIT_BUREAU_YEAR  \n",
       "0                         1.0  \n",
       "1                         0.0  \n",
       "2                         0.0  \n",
       "3                         NaN  \n",
       "4                         0.0  \n",
       "\n",
       "[5 rows x 122 columns]"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# All train application data\n",
    "application_train = pd.read_csv(\"data/application_train.csv\")\n",
    "\n",
    "# Train data without rows with any null value column\n",
    "application_train_nmv = application_train.dropna() \n",
    "\n",
    "application_train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total train rows: 307511\n",
      "Rows tagged as possible default: 24825 (8.07% of the rows)\n",
      "\n",
      "Total train rows (without missing values): 8602\n",
      "Rows tagged as possible default: 526 (6.11% of the rows)\n"
     ]
    },
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "C:\\Users\\user\\Anaconda3\\lib\\site-packages\\ipykernel_launcher.py:12: UserWarning: Boolean Series key will be reindexed to match DataFrame index.\n",
      "  if sys.path[0] == '':\n"
     ]
    }
   ],
   "source": [
    "# Full train data\n",
    "print(\"Total train rows: {}\".format(application_train.shape[0]))\n",
    "is_ok = application_train['TARGET'] == 1\n",
    "application_train_ok = application_train[is_ok]\n",
    "ok_percent = (application_train_ok.shape[0] / application_train.shape[0]) * 100\n",
    "print(\"Rows tagged as possible default: {} ({:.2f}% of the rows)\".format(application_train_ok.shape[0], ok_percent))\n",
    "\n",
    "# No missing values data\n",
    "print()\n",
    "print(\"Total train rows (without missing values): {}\".format(application_train_nmv.shape[0]))\n",
    "is_ok_nmv = application_train_nmv['TARGET'] == 1\n",
    "application_train_ok_nmv = application_train_nmv[is_ok]\n",
    "ok_percent_nmv = (application_train_ok_nmv.shape[0] / application_train_nmv.shape[0]) * 100\n",
    "print(\"Rows tagged as possible default: {} ({:.2f}% of the rows)\".format(application_train_ok_nmv.shape[0], ok_percent_nmv))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['AMT_ANNUITY', 'AMT_GOODS_PRICE', 'NAME_TYPE_SUITE', 'OWN_CAR_AGE',\n",
       "       'OCCUPATION_TYPE', 'CNT_FAM_MEMBERS', 'EXT_SOURCE_1', 'EXT_SOURCE_2',\n",
       "       'EXT_SOURCE_3', 'APARTMENTS_AVG', 'BASEMENTAREA_AVG',\n",
       "       'YEARS_BEGINEXPLUATATION_AVG', 'YEARS_BUILD_AVG', 'COMMONAREA_AVG',\n",
       "       'ELEVATORS_AVG', 'ENTRANCES_AVG', 'FLOORSMAX_AVG', 'FLOORSMIN_AVG',\n",
       "       'LANDAREA_AVG', 'LIVINGAPARTMENTS_AVG', 'LIVINGAREA_AVG',\n",
       "       'NONLIVINGAPARTMENTS_AVG', 'NONLIVINGAREA_AVG', 'APARTMENTS_MODE',\n",
       "       'BASEMENTAREA_MODE', 'YEARS_BEGINEXPLUATATION_MODE', 'YEARS_BUILD_MODE',\n",
       "       'COMMONAREA_MODE', 'ELEVATORS_MODE', 'ENTRANCES_MODE', 'FLOORSMAX_MODE',\n",
       "       'FLOORSMIN_MODE', 'LANDAREA_MODE', 'LIVINGAPARTMENTS_MODE',\n",
       "       'LIVINGAREA_MODE', 'NONLIVINGAPARTMENTS_MODE', 'NONLIVINGAREA_MODE',\n",
       "       'APARTMENTS_MEDI', 'BASEMENTAREA_MEDI', 'YEARS_BEGINEXPLUATATION_MEDI',\n",
       "       'YEARS_BUILD_MEDI', 'COMMONAREA_MEDI', 'ELEVATORS_MEDI',\n",
       "       'ENTRANCES_MEDI', 'FLOORSMAX_MEDI', 'FLOORSMIN_MEDI', 'LANDAREA_MEDI',\n",
       "       'LIVINGAPARTMENTS_MEDI', 'LIVINGAREA_MEDI', 'NONLIVINGAPARTMENTS_MEDI',\n",
       "       'NONLIVINGAREA_MEDI', 'FONDKAPREMONT_MODE', 'HOUSETYPE_MODE',\n",
       "       'TOTALAREA_MODE', 'WALLSMATERIAL_MODE', 'EMERGENCYSTATE_MODE',\n",
       "       'OBS_30_CNT_SOCIAL_CIRCLE', 'DEF_30_CNT_SOCIAL_CIRCLE',\n",
       "       'OBS_60_CNT_SOCIAL_CIRCLE', 'DEF_60_CNT_SOCIAL_CIRCLE',\n",
       "       'DAYS_LAST_PHONE_CHANGE', 'AMT_REQ_CREDIT_BUREAU_HOUR',\n",
       "       'AMT_REQ_CREDIT_BUREAU_DAY', 'AMT_REQ_CREDIT_BUREAU_WEEK',\n",
       "       'AMT_REQ_CREDIT_BUREAU_MON', 'AMT_REQ_CREDIT_BUREAU_QRT',\n",
       "       'AMT_REQ_CREDIT_BUREAU_YEAR'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "application_train.columns[application_train.isnull().any()]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Note:\n",
    "For the Application Training table there are too many rows with missing values, so I will not exclude this rows.<p>\n",
    "My strategy will be for each column to analyze and take the proper action on missing values, skewed data, categorical data, etc. </p>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train = application_train"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TABLE application_train.csv\n",
    "### SK_ID_CURR\n",
    "ID of loan in our sample<br>\n",
    "This column will be removed after the analysis, because it is needed for table merging."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### TARGET\n",
    "Is the target variable and no treatment is necessary (1 - client with payment difficulties, 0 - all other cases)."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NAME_CONTRACT_TYPE\n",
    "Identification if loan is cash or revolving. <br>\n",
    "Categorial."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in NAME_CONTRACT_TYPE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in NAME_CONTRACT_TYPE: {}\".format(\n",
    "    processed_app_train[processed_app_train.NAME_CONTRACT_TYPE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for NAME_CONTRACT_TYPE:  ['Cash loans' 'Revolving loans']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for NAME_CONTRACT_TYPE: \", processed_app_train['NAME_CONTRACT_TYPE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['NAME_CONTRACT_TYPE'], prefix='CTYPE', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('NAME_CONTRACT_TYPE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CODE_GENDER\n",
    "Gender of the client. <br>\n",
    "Categorial."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in CODE_GENDER: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in CODE_GENDER: {}\".format(\n",
    "    processed_app_train[processed_app_train.CODE_GENDER.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for CODE_GENDER:  ['M' 'F' 'XNA']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for CODE_GENDER: \", processed_app_train['CODE_GENDER'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['CODE_GENDER'], prefix='GEND', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('CODE_GENDER',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_OWN_CAR\n",
    "Flag if the client owns a car. <br>\n",
    "Categorial."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_OWN_CAR: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_OWN_CAR: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_OWN_CAR.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for FLAG_OWN_CAR:  ['N' 'Y']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for FLAG_OWN_CAR: \", processed_app_train['FLAG_OWN_CAR'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Value replacing (Y->1 and N->0)\n",
    "processed_app_train['FLAG_OWN_CAR'] = processed_app_train['FLAG_OWN_CAR'].replace('Y', 1)\n",
    "processed_app_train['FLAG_OWN_CAR'] = processed_app_train['FLAG_OWN_CAR'].replace('N', 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_OWN_REALTY\n",
    "Flag if client owns a house or flat. <br>\n",
    "Categorial."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_OWN_REALTY: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_OWN_REALTY: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_OWN_REALTY.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for FLAG_OWN_REALTY:  ['Y' 'N']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for FLAG_OWN_REALTY: \", processed_app_train['FLAG_OWN_REALTY'].unique()) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Value replacement (Y->1 and N->0)\n",
    "processed_app_train['FLAG_OWN_REALTY'] = processed_app_train['FLAG_OWN_REALTY'].replace('Y', 1)\n",
    "processed_app_train['FLAG_OWN_REALTY'] = processed_app_train['FLAG_OWN_REALTY'].replace('N', 0)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CNT_CHILDREN\n",
    "Number of children the client has."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in CNT_CHILDREN: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in CNT_CHILDREN: {}\".format(\n",
    "    processed_app_train[processed_app_train.CNT_CHILDREN.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c0475d1d0>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['CNT_CHILDREN'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['CNT_CHILDREN'] = ((processed_app_train['CNT_CHILDREN'] -\n",
    "                                        processed_app_train['CNT_CHILDREN'].min())/\n",
    "                                       (processed_app_train['CNT_CHILDREN'].max() - \n",
    "                                        processed_app_train['CNT_CHILDREN'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_INCOME_TOTAL\n",
    "Income of the client."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in AMT_INCOME_TOTAL: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in AMT_INCOME_TOTAL: {}\".format(\n",
    "    processed_app_train[processed_app_train.AMT_INCOME_TOTAL.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c04a8c5f8>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['AMT_INCOME_TOTAL'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 391.5596541041876\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['AMT_INCOME_TOTAL'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c04b77dd8>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['AMT_INCOME_TOTAL'].apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['AMT_INCOME_TOTAL'] = np.log(processed_app_train.AMT_INCOME_TOTAL)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_INCOME_TOTAL'] = ((processed_app_train['AMT_INCOME_TOTAL'] -\n",
    "                                            processed_app_train['AMT_INCOME_TOTAL'].min())/\n",
    "                                           (processed_app_train['AMT_INCOME_TOTAL'].max() - \n",
    "                                            processed_app_train['AMT_INCOME_TOTAL'].min()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['AMT_INCOME_TOTAL'] = processed_app_train['AMT_INCOME_TOTAL'].replace(0, 0.0001)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_CREDIT\n",
    "Credit amount of the loan."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in AMT_CREDIT: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in AMT_CREDIT: {}\".format(\n",
    "    processed_app_train[processed_app_train.AMT_CREDIT.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c04db1e48>"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['AMT_CREDIT'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 1.2347784971349398\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['AMT_CREDIT'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c04b70b00>"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAD8CAYAAACcjGjIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAFSRJREFUeJzt3X+s5XWd3/HnawexFLoLLnJLYdIh66QRnd1ZvQES/+hVWxiwWbDVBkt0UJqxG0h302nqsCbFiiSYDUtiV0nGMAG3riPZXcNEx+LEckM2WRRYkWFEyy3OyoUJxA6io63m6rt/nO+UM/M5d+65P88Z5vlITu457/P5fs/7+5l77+t+f5wzqSokSer3a6NuQJI0fgwHSVLDcJAkNQwHSVLDcJAkNQwHSVLDcJAkNQwHSVLDcJAkNU4bdQNLde6559aGDRtG3QY//elPOfPMM0fdxthwPo7lfLzCuTjWqObjscce+2FVvX6hcSdtOGzYsIFHH3101G0wPT3N1NTUqNsYG87HsZyPVzgXxxrVfCT5u2HGeVhJktQwHCRJDcNBktQwHCRJDcNBktQwHCRJjQXDIcnfS/LNJN9OciDJf+nqFyX5RpKnk3wxyeld/bXd45nu+Q1967q5q38vyRV99S1dbSbJjpXfTEnSYgyz5/Bz4B1V9TvAZmBLksuATwJ3VtVG4CXghm78DcBLVfUG4M5uHEkuBq4F3gRsAT6TZF2SdcCngSuBi4H3dWMlSSOyYDhUz5Hu4Wu6WwHvAP6iq98LXNPdv7p7TPf8O5Okq++uqp9X1feBGeCS7jZTVc9U1S+A3d1YSdKIDPUO6e6v+8eAN9D7K/9/AT+qqrluyCxwQXf/AuBZgKqaS/Iy8Jtd/eG+1fYv8+xx9UsXvSXSSWbDjq8MNe7g7e9a5U6k1lDhUFW/BDYnORv4EvDGQcO6r5nnufnqg/ZeakCNJNuAbQATExNMT0+fuPE1cOTIkbHoY1w4H8c60Xxs3zQ3sH68V8t8+r1xrHGfj0V9tlJV/SjJNHAZcHaS07q9hwuB57ths8B6YDbJacBvAIf76kf1LzNf/fjX3wnsBJicnKxx+JwWPy/mWM7HsU40H9cPu+dw3eDlTzZ+bxxr3OdjmKuVXt/tMZDkDOCfAU8BDwLv6YZtBe7v7u/pHtM9/z+qqrr6td3VTBcBG4FvAo8AG7urn06nd9J6z0psnCRpaYbZczgfuLc77/BrwH1V9eUk3wF2J/kE8C3g7m783cCfJZmht8dwLUBVHUhyH/AdYA64sTtcRZKbgAeAdcCuqjqwYlsoSVq0BcOhqp4AfndA/Rl6VxodX/+/wHvnWddtwG0D6nuBvUP0K0laA75DWpLUMBwkSQ3DQZLUMBwkSQ3DQZLUMBwkSQ3DQZLUMBwkSQ3DQZLUMBwkSQ3DQZLUMBwkSQ3DQZLUMBwkSY1F/U9wksab/y+1Vop7DpKkhnsO0inIPQwtxHDQScdfbNLq87CSJKlhOEiSGoaDJKlhOEiSGoaDJKlhOEiSGoaDJKlhOEiSGguGQ5L1SR5M8lSSA0n+oKt/LMlzSR7vblf1LXNzkpkk30tyRV99S1ebSbKjr35Rkm8keTrJF5OcvtIbKkka3jB7DnPA9qp6I3AZcGOSi7vn7qyqzd1tL0D33LXAm4AtwGeSrEuyDvg0cCVwMfC+vvV8slvXRuAl4IYV2j5J0hIs+PEZVXUIONTd/0mSp4ALTrDI1cDuqvo58P0kM8Al3XMzVfUMQJLdwNXd+t4B/JtuzL3Ax4C7Fr850uj1f7zH9k1zXD/kx31I4yRVNfzgZAPwEPBm4D8A1wM/Bh6lt3fxUpI/BR6uqv/WLXM38NVuFVuq6t929fcDl9ILgoer6g1dfT3w1ap684DX3wZsA5iYmHjr7t27F7e1q+DIkSOcddZZo25jbKzFfOx/7uWhxm264DdWtY/59Pc3cQa88H+Wt77FbMewc7Mar70Qf1aONar5ePvb3/5YVU0uNG7oD95Lchbwl8AfVtWPk9wF3ApU9/UO4ENABixeDD6EVScY3xardgI7ASYnJ2tqamrY9lfN9PQ049DHuFiL+Rj2L/GD161uH/O5/rg9hzv2L+/zLRezHSu9l7KSc+jPyrHGfT6G+q5N8hp6wfD5qvorgKp6oe/5zwJf7h7OAuv7Fr8QeL67P6j+Q+DsJKdV1dxx4yVJIzDM1UoB7gaeqqo/6auf3zfs3cCT3f09wLVJXpvkImAj8E3gEWBjd2XS6fROWu+p3nGtB4H3dMtvBe5f3mZJkpZjmD2HtwHvB/Ynebyr/RG9q4020zsEdBD4MEBVHUhyH/Adelc63VhVvwRIchPwALAO2FVVB7r1fQTYneQTwLfohZEkaUSGuVrprxl8XmDvCZa5DbhtQH3voOW6K5guOb4uSRoN3yEtSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWoYDpKkhuEgSWosGA5J1id5MMlTSQ4k+YOu/rok+5I83X09p6snyaeSzCR5Islb+ta1tRv/dJKtffW3JtnfLfOpJFmNjZUkDWeYPYc5YHtVvRG4DLgxycXADuDrVbUR+Hr3GOBKYGN32wbcBb0wAW4BLgUuAW45GijdmG19y21Z/qZJkpZqwXCoqkNV9bfd/Z8ATwEXAFcD93bD7gWu6e5fDXyueh4Gzk5yPnAFsK+qDlfVS8A+YEv33K9X1d9UVQGf61uXJGkEFnXOIckG4HeBbwATVXUIegECnNcNuwB4tm+x2a52ovrsgLokaUROG3ZgkrOAvwT+sKp+fILTAoOeqCXUB/Wwjd7hJyYmJpienl6g69V35MiRsehjXKzFfGzfNDfUuFH9u/T3N3HG8P3OZzHbsdzXWs5rL8SflWON+3wMFQ5JXkMvGD5fVX/VlV9Icn5VHeoODb3Y1WeB9X2LXwg839WnjqtPd/ULB4xvVNVOYCfA5ORkTU1NDRq2pqanpxmHPsbFWszH9Tu+MtS4g9etbh/z6e9v+6Y57tg/9N9gAy1mO4adm9V47YX4s3KscZ+PYa5WCnA38FRV/UnfU3uAo1ccbQXu76t/oLtq6TLg5e6w0wPA5UnO6U5EXw480D33kySXda/1gb51SZJGYJg/ad4GvB/Yn+TxrvZHwO3AfUluAH4AvLd7bi9wFTAD/Az4IEBVHU5yK/BIN+7jVXW4u//7wD3AGcBXu5u0LBuG3cO4/V2r3Il08lkwHKrqrxl8XgDgnQPGF3DjPOvaBewaUH8UePNCvUiS1obvkJYkNQwHSVLDcJAkNQwHSVLDcJAkNQwHSVLDcJAkNQwHSVLDcJAkNQwHSVLDcJAkNQwHSVLDcJAkNQwHSVLDcJAkNQwHSVJjef+5raRVN+z/aCetJPccJEkNw0GS1DAcJEkNw0GS1DAcJEkNr1aStGzDXFG1fdMcU6vfilaI4aCx4OWa48l/l1OXh5UkSY0FwyHJriQvJnmyr/axJM8leby7XdX33M1JZpJ8L8kVffUtXW0myY6++kVJvpHk6SRfTHL6Sm6gJGnxhtlzuAfYMqB+Z1Vt7m57AZJcDFwLvKlb5jNJ1iVZB3wauBK4GHhfNxbgk926NgIvATcsZ4MkScu3YDhU1UPA4SHXdzWwu6p+XlXfB2aAS7rbTFU9U1W/AHYDVycJ8A7gL7rl7wWuWeQ2SJJW2HLOOdyU5InusNM5Xe0C4Nm+MbNdbb76bwI/qqq54+qSpBFa6tVKdwG3AtV9vQP4EJABY4vBIVQnGD9Qkm3ANoCJiQmmp6cX1fRqOHLkyFj0MS6WOh/bN80tPGiVrPS/X/+2TJwx2m0bJxNnrPxcn8zG/XfHksKhql44ej/JZ4Evdw9ngfV9Qy8Enu/uD6r/EDg7yWnd3kP/+EGvuxPYCTA5OVlTU1NLaX9FTU9PMw59jIulzsf1I7xk8uB1Uyu6vv5t2b5pjjv2e8U49ObiX/uz8v+N+++OJR1WSnJ+38N3A0evZNoDXJvktUkuAjYC3wQeATZ2VyadTu+k9Z6qKuBB4D3d8luB+5fSkyRp5Sz4J02SLwBTwLlJZoFbgKkkm+kdAjoIfBigqg4kuQ/4DjAH3FhVv+zWcxPwALAO2FVVB7qX+AiwO8kngG8Bd6/Y1kmSlmTBcKiq9w0oz/sLvKpuA24bUN8L7B1Qf4be1UySpDHhO6QlSQ3PlElaM8N+VtPB29+1yp1oIe45SJIahoMkqWE4SJIannPQKc/j4FLLPQdJUsNwkCQ1DAdJUsNwkCQ1DAdJUsNwkCQ1DAdJUsNwkCQ1DAdJUsNwkCQ1DAdJUsNwkCQ1DAdJUsNwkCQ1/MhuraphPw5b0nhxz0GS1DAcJEkNw0GS1DAcJEkNw0GS1FgwHJLsSvJikif7aq9Lsi/J093Xc7p6knwqyUySJ5K8pW+Zrd34p5Ns7au/Ncn+bplPJclKb6QkaXGGuZT1HuBPgc/11XYAX6+q25Ps6B5/BLgS2NjdLgXuAi5N8jrgFmASKOCxJHuq6qVuzDbgYWAvsAX46vI3TVpZXparU8mCew5V9RBw+Ljy1cC93f17gWv66p+rnoeBs5OcD1wB7Kuqw10g7AO2dM/9elX9TVUVvQC6BknSSC31TXATVXUIoKoOJTmvq18APNs3brarnag+O6A+UJJt9PYymJiYYHp6eontr5wjR46MRR/j4vj52L5pbnTNjIGJM5yDoxYzF6fCz9S4/+5Y6XdIDzpfUEuoD1RVO4GdAJOTkzU1NbWEFlfW9PQ049DHuDh+Pq4/xQ/FbN80xx37/SACWNxcHLxuanWbGQPj/rtjqVcrvdAdEqL7+mJXnwXW9427EHh+gfqFA+qSpBFaajjsAY5ecbQVuL+v/oHuqqXLgJe7w08PAJcnOae7suly4IHuuZ8kuay7SukDfeuSJI3Igvt4Sb4ATAHnJpmld9XR7cB9SW4AfgC8txu+F7gKmAF+BnwQoKoOJ7kVeKQb9/GqOnqS+/fpXRF1Br2rlLxSSZJGbMFwqKr3zfPUOweMLeDGedazC9g1oP4o8OaF+pAkrR3fIS1JahgOkqSG4SBJahgOkqSG4SBJahgOkqSG4SBJahgOkqSG4SBJahgOkqSG4SBJahgOkqSG4SBJahgOkqSG/3+hjrFhyP/W8+Dt71rlTiSNknsOkqSG4SBJanhYSUsy3+Gn7ZvmuH7IQ1OSxpfhIGnseO5r9DysJElqGA6SpIbhIElqGA6SpIbhIElqGA6SpMaywiHJwST7kzye5NGu9rok+5I83X09p6snyaeSzCR5Islb+taztRv/dJKty9skSdJyrcSew9uranNVTXaPdwBfr6qNwNe7xwBXAhu72zbgLuiFCXALcClwCXDL0UCRJI3GahxWuhq4t7t/L3BNX/1z1fMwcHaS84ErgH1VdbiqXgL2AVtWoS9J0pCWGw4FfC3JY0m2dbWJqjoE0H09r6tfADzbt+xsV5uvLkkakeV+fMbbqur5JOcB+5J89wRjM6BWJ6i3K+gF0DaAiYkJpqenF9nuyjty5MhY9LFStm+aW9byE2csfx2vJs7HK1ZjLk7mn71x/92xrHCoque7ry8m+RK9cwYvJDm/qg51h41e7IbPAuv7Fr8QeL6rTx1Xn57n9XYCOwEmJydrampq0LA1NT09zTj0sVKW+6F52zfNccd+P7LrKOfjFasxFwevm1rR9a2lcf/dseTDSknOTPIPjt4HLgeeBPYAR6842grc393fA3ygu2rpMuDl7rDTA8DlSc7pTkRf3tUkSSOynBifAL6U5Oh6/ryq/nuSR4D7ktwA/AB4bzd+L3AVMAP8DPggQFUdTnIr8Eg37uNVdXgZfUmSlmnJ4VBVzwC/M6D+v4F3DqgXcOM869oF7FpqL5KkleU7pCVJDcNBktQwHCRJDcNBktQwHCRJDd+dI+mktWERb9o8ePu7VrGTVx/3HCRJDcNBktTwsNIpYjG735JkOJzk/KUvaTV4WEmS1DAcJEkNw0GS1DAcJEkNw0GS1DAcJEkNL2WVdEoY9rJvP2ajxz0HSVLDcJAkNTystIb8BElJJwvDYUz5sRiSRumUDAdPTEnSiXnOQZLUOCX3HCRpPh5Z6DEcVoDnByS92ozNYaUkW5J8L8lMkh2j7keSTmVjseeQZB3waeCfA7PAI0n2VNV3RtnXMHsE2zfNMSbTKEkrZlz2HC4BZqrqmar6BbAbuHrEPUnSKWtc/uS9AHi27/EscOmIepGkBb3aT1yPSzhkQK2aQck2YFv38EiS761qV0P493Au8MNR9zEunI9jOR+vOFXnIp+c96lRzcc/HmbQuITDLLC+7/GFwPPHD6qqncDOtWpqGEkerarJUfcxLpyPYzkfr3AujjXu8zEu5xweATYmuSjJ6cC1wJ4R9yRJp6yx2HOoqrkkNwEPAOuAXVV1YMRtSdIpayzCAaCq9gJ7R93HEozVYa4x4Hwcy/l4hXNxrLGej1Q1530lSae4cTnnIEkaI4bDIiTZleTFJE/21d6b5ECSXyUZ2ysPVsM88/HHSb6b5IkkX0py9ih7XCvzzMWt3Tw8nuRrSf7RKHtcS4Pmo++5/5ikkpw7it5GYZ7vj48lea77/ng8yVWj7PF4hsPi3ANsOa72JPAvgYfWvJvRu4d2PvYBb66q3wb+J3DzWjc1IvfQzsUfV9VvV9Vm4MvAf17zrkbnHtr5IMl6eh+T84O1bmjE7mHAfAB3VtXm7jZW51wNh0WoqoeAw8fVnqqqkb8ZbxTmmY+vVdVc9/Bheu9ZedWbZy5+3PfwTAa8sfPVatB8dO4E/hOn0FzACedjbBkOWk0fAr466iZGKcltSZ4FruPU2nNoJPk94Lmq+vaoexkjN3WHHnclOWfUzfQzHLQqknwUmAM+P+peRqmqPlpV6+nNw02j7mdUkvx94KOc4gF5nLuA3wI2A4eAO0bbzrEMB624JFuBfwFcV14rfdSfA/9q1E2M0G8BFwHfTnKQ3uHGv03yD0fa1QhV1QtV9cuq+hXwWXqfTj02xuZNcHp1SLIF+AjwT6vqZ6PuZ5SSbKyqp7uHvwd8d5T9jFJV7QfOO/q4C4jJqjrlPojvqCTnV9Wh7uG76V3cMjYMh0VI8gVgCjg3ySxwC72TTP8VeD3wlSSPV9UVo+ty7cwzHzcDrwX2JQF4uKr+3ciaXCPzzMVVSf4J8Cvg74BX/TwcNWg+quru0XY1OvN8f0wl2Uzv5PxB4MMja3AA3yEtSWp4zkGS1DAcJEkNw0GS1DAcJEkNw0GS1DAcJEkNw0GS1DAcJEmN/wdG3qPZsGn8sgAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['AMT_CREDIT'].apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['AMT_CREDIT'] = np.log(processed_app_train.AMT_CREDIT)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_CREDIT'] = ((processed_app_train['AMT_CREDIT'] -\n",
    "                                      processed_app_train['AMT_CREDIT'].min())/\n",
    "                                     (processed_app_train['AMT_CREDIT'].max() - \n",
    "                                      processed_app_train['AMT_CREDIT'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_ANNUITY\n",
    "Loan annuity."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in AMT_ANNUITY: 12\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in AMT_ANNUITY: {}\".format(\n",
    "    processed_app_train[processed_app_train.AMT_ANNUITY.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Few rows with nulls -> delete rows\n",
    "processed_app_train = processed_app_train[processed_app_train.AMT_ANNUITY.notnull()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c07acff28>"
      ]
     },
     "execution_count": 36,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['AMT_ANNUITY'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 1.5797773638612504\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['AMT_ANNUITY'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c07771898>"
      ]
     },
     "execution_count": 38,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['AMT_ANNUITY'].apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['AMT_ANNUITY'] = np.log(processed_app_train.AMT_ANNUITY)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_ANNUITY'] = ((processed_app_train['AMT_ANNUITY'] -\n",
    "                                      processed_app_train['AMT_ANNUITY'].min())/\n",
    "                                     (processed_app_train['AMT_ANNUITY'].max() - \n",
    "                                      processed_app_train['AMT_ANNUITY'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_GOODS_PRICE\n",
    "For consumer loans it is the price of the goods for which the loan is given."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in AMT_GOODS_PRICE: 278\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in AMT_GOODS_PRICE: {}\".format(\n",
    "    processed_app_train[processed_app_train.AMT_GOODS_PRICE.isnull()].shape[0]))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Few rows with nulls -> delete rows\n",
    "processed_app_train = processed_app_train[processed_app_train.AMT_GOODS_PRICE.notnull()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c04aacb70>"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['AMT_GOODS_PRICE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 1.3489957352316093\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['AMT_GOODS_PRICE'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c0765b4e0>"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['AMT_GOODS_PRICE'].apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['AMT_GOODS_PRICE'] = np.log(processed_app_train.AMT_GOODS_PRICE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_GOODS_PRICE'] = ((processed_app_train['AMT_GOODS_PRICE'] -\n",
    "                                      processed_app_train['AMT_GOODS_PRICE'].min())/\n",
    "                                     (processed_app_train['AMT_GOODS_PRICE'].max() - \n",
    "                                      processed_app_train['AMT_GOODS_PRICE'].min()))\n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NAME_TYPE_SUITE\n",
    "Who was accompanying client when he was applying for the loan. <br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in NAME_TYPE_SUITE: 1014\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in NAME_TYPE_SUITE: {}\".format(\n",
    "    processed_app_train[processed_app_train.NAME_TYPE_SUITE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for NAME_TYPE_SUITE:  ['Unaccompanied' 'Family' 'Spouse, partner' 'Children' 'Other_A' nan\n",
      " 'Other_B' 'Group of people']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for NAME_TYPE_SUITE: \", processed_app_train['NAME_TYPE_SUITE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['NAME_TYPE_SUITE'].fillna(value='Other', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['NAME_TYPE_SUITE'], prefix='NTS', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('NAME_TYPE_SUITE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NAME_INCOME_TYPE\n",
    "Clients income type (businessman, working, maternity leave,…).<br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in NAME_INCOME_TYPE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in NAME_INCOME_TYPE: {}\".format(\n",
    "    processed_app_train[processed_app_train.NAME_INCOME_TYPE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for NAME_INCOME_TYPE:  ['Working' 'State servant' 'Commercial associate' 'Pensioner' 'Unemployed'\n",
      " 'Student' 'Businessman' 'Maternity leave']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for NAME_INCOME_TYPE: \", processed_app_train['NAME_INCOME_TYPE'].unique())\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['NAME_INCOME_TYPE'], prefix='INCTYP', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('NAME_INCOME_TYPE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NAME_EDUCATION_TYPE\n",
    "Level of highest education the client achieved. <br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in NAME_EDUCATION_TYPE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in NAME_EDUCATION_TYPE: {}\".format(\n",
    "    processed_app_train[processed_app_train.NAME_EDUCATION_TYPE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for NAME_EDUCATION_TYPE:  ['Secondary / secondary special' 'Higher education' 'Incomplete higher'\n",
      " 'Lower secondary' 'Academic degree']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for NAME_EDUCATION_TYPE: \", processed_app_train['NAME_EDUCATION_TYPE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['NAME_EDUCATION_TYPE'] = processed_app_train['NAME_EDUCATION_TYPE'].replace(\n",
    "    'Secondary / secondary special', 'Secondary special')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['NAME_EDUCATION_TYPE'], prefix='EDU', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('NAME_EDUCATION_TYPE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NAME_FAMILY_STATUS\n",
    "Family status of the client. <br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in NAME_FAMILY_STATUS: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in NAME_FAMILY_STATUS: {}\".format(\n",
    "    processed_app_train[processed_app_train.NAME_FAMILY_STATUS.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for NAME_FAMILY_STATUS:  ['Single / not married' 'Married' 'Civil marriage' 'Widow' 'Separated']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for NAME_FAMILY_STATUS: \", processed_app_train['NAME_FAMILY_STATUS'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['NAME_FAMILY_STATUS'] = processed_app_train['NAME_FAMILY_STATUS'].replace(\n",
    "    'Single / not married', 'Single')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['NAME_FAMILY_STATUS'], prefix='FAM', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('NAME_FAMILY_STATUS',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### NAME_HOUSING_TYPE\n",
    "What is the housing situation of the client (renting, living with parents, ...). <br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in NAME_HOUSING_TYPE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in NAME_HOUSING_TYPE: {}\".format(\n",
    "    processed_app_train[processed_app_train.NAME_HOUSING_TYPE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for NAME_HOUSING_TYPE:  ['House / apartment' 'Rented apartment' 'With parents'\n",
      " 'Municipal apartment' 'Office apartment' 'Co-op apartment']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for NAME_HOUSING_TYPE: \", processed_app_train['NAME_HOUSING_TYPE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['NAME_HOUSING_TYPE'] = processed_app_train['NAME_HOUSING_TYPE'].replace(\n",
    "    'House / apartment', 'House')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['NAME_HOUSING_TYPE'], prefix='HOUST', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('NAME_HOUSING_TYPE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REGION_POPULATION_RELATIVE\n",
    "Normalized population of region where client lives (higher number means the client lives in more populated region).\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in REGION_POPULATION_RELATIVE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in REGION_POPULATION_RELATIVE: {}\".format(\n",
    "    processed_app_train[processed_app_train.REGION_POPULATION_RELATIVE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c077350f0>"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['REGION_POPULATION_RELATIVE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['REGION_POPULATION_RELATIVE'] = ((processed_app_train['REGION_POPULATION_RELATIVE'] -\n",
    "                                      processed_app_train['REGION_POPULATION_RELATIVE'].min())/\n",
    "                                     (processed_app_train['REGION_POPULATION_RELATIVE'].max() - \n",
    "                                      processed_app_train['REGION_POPULATION_RELATIVE'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DAYS_BIRTH\n",
    "Client's age in days at the time of application."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in DAYS_BIRTH: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in DAYS_BIRTH: {}\".format(\n",
    "    processed_app_train[processed_app_train.DAYS_BIRTH.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c0797d7f0>"
      ]
     },
     "execution_count": 71,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['DAYS_BIRTH'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['DAYS_BIRTH'] = ((processed_app_train['DAYS_BIRTH'] -\n",
    "                                      processed_app_train['DAYS_BIRTH'].min())/\n",
    "                                     (processed_app_train['DAYS_BIRTH'].max() - \n",
    "                                      processed_app_train['DAYS_BIRTH'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DAYS_EMPLOYED\n",
    "How many days before the application the person started current employment."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in DAYS_EMPLOYED: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in DAYS_EMPLOYED: {}\".format(\n",
    "    processed_app_train[processed_app_train.DAYS_EMPLOYED.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c07a2d080>"
      ]
     },
     "execution_count": 74,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['DAYS_EMPLOYED'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['DAYS_EMPLOYED'] = ((processed_app_train['DAYS_EMPLOYED'] -\n",
    "                                      processed_app_train['DAYS_EMPLOYED'].min())/\n",
    "                                     (processed_app_train['DAYS_EMPLOYED'].max() - \n",
    "                                      processed_app_train['DAYS_EMPLOYED'].min()))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DAYS_REGISTRATION\n",
    "How many days before the application did client change his registration."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in DAYS_REGISTRATION: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in DAYS_REGISTRATION: {}\".format(\n",
    "    processed_app_train[processed_app_train.DAYS_REGISTRATION.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c07b2d4e0>"
      ]
     },
     "execution_count": 77,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['DAYS_REGISTRATION'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['DAYS_REGISTRATION'] = ((processed_app_train['DAYS_REGISTRATION'] -\n",
    "                                      processed_app_train['DAYS_REGISTRATION'].min())/\n",
    "                                     (processed_app_train['DAYS_REGISTRATION'].max() - \n",
    "                                      processed_app_train['DAYS_REGISTRATION'].min()))\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DAYS_ID_PUBLISH\n",
    "How many days before the application did client change the identity document with which he applied for the loan."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in DAYS_ID_PUBLISH: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in DAYS_ID_PUBLISH: {}\".format(\n",
    "    processed_app_train[processed_app_train.DAYS_ID_PUBLISH.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c079fb2b0>"
      ]
     },
     "execution_count": 80,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['DAYS_ID_PUBLISH'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['DAYS_ID_PUBLISH'] = ((processed_app_train['DAYS_ID_PUBLISH'] -\n",
    "                                      processed_app_train['DAYS_ID_PUBLISH'].min())/\n",
    "                                     (processed_app_train['DAYS_ID_PUBLISH'].max() - \n",
    "                                      processed_app_train['DAYS_ID_PUBLISH'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OWN_CAR_AGE\n",
    "Age of client's car."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in OWN_CAR_AGE: 202642\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in OWN_CAR_AGE: {}\".format(\n",
    "    processed_app_train[processed_app_train.OWN_CAR_AGE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19936e48>"
      ]
     },
     "execution_count": 83,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['OWN_CAR_AGE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The spike for cars above 60 years may be related to antiques. But after a certain age it is difficult to know if we are talking about an antique (expensive car) or just a really old one (very inexpensive). Cars above 40 years may be considered an antique. So, let's see how many cars are in this class:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Applicants with car with more then 40 years: 3526\n"
     ]
    }
   ],
   "source": [
    "print(\"Applicants with car with more then 40 years: {}\".format(\n",
    "    processed_app_train[processed_app_train['OWN_CAR_AGE'] > 40].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [],
   "source": [
    "# delete rows with cars above 40 years\n",
    "processed_app_train.drop(processed_app_train[processed_app_train['OWN_CAR_AGE'] > 40].index , inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {},
   "outputs": [],
   "source": [
    "# I will fill the missing values of the age with the mean value of the remaing ages\n",
    "processed_app_train['OWN_CAR_AGE'].fillna(value=processed_app_train['OWN_CAR_AGE'].mean(), inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c199efe80>"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['OWN_CAR_AGE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['OWN_CAR_AGE'] = ((processed_app_train['OWN_CAR_AGE'] -\n",
    "                                      processed_app_train['OWN_CAR_AGE'].min())/\n",
    "                                     (processed_app_train['OWN_CAR_AGE'].max() - \n",
    "                                      processed_app_train['OWN_CAR_AGE'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_MOBIL\n",
    "Did client provide mobile phone (1=YES, 0=NO)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_MOBIL: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_MOBIL: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_MOBIL.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_EMP_PHONE\n",
    "Did client provide work phone (1=YES, 0=NO)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_EMP_PHONE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_EMP_PHONE: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_EMP_PHONE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_WORK_PHONE\n",
    "Did client provide home phone (1=YES, 0=NO)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_WORK_PHONE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_WORK_PHONE: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_WORK_PHONE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_CONT_MOBILE\n",
    "Was mobile phone reachable (1=YES, 0=NO)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_CONT_MOBILE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_CONT_MOBILE: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_CONT_MOBILE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_PHONE\n",
    "Did client provide home phone (1=YES, 0=NO)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_PHONE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_PHONE: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_PHONE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_EMAIL\n",
    "Did client provide email (1=YES, 0=NO)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FLAG_EMAIL: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FLAG_EMAIL: {}\".format(\n",
    "    processed_app_train[processed_app_train.FLAG_EMAIL.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OCCUPATION_TYPE\n",
    "What kind of occupation does the client have.\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in OCCUPATION_TYPE: 95529\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in OCCUPATION_TYPE: {}\".format(\n",
    "    processed_app_train[processed_app_train.OCCUPATION_TYPE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for OCCUPATION_TYPE:  ['Laborers' 'Core staff' 'Accountants' 'Managers' nan 'Drivers'\n",
      " 'Sales staff' 'Cleaning staff' 'Cooking staff' 'Private service staff'\n",
      " 'Medicine staff' 'Security staff' 'High skill tech staff'\n",
      " 'Waiters/barmen staff' 'Low-skill Laborers' 'Realty agents' 'Secretaries'\n",
      " 'IT staff' 'HR staff']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for OCCUPATION_TYPE: \", processed_app_train['OCCUPATION_TYPE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['OCCUPATION_TYPE'].fillna(value='None', inplace=True)\n",
    "processed_app_train['OCCUPATION_TYPE'] = processed_app_train['OCCUPATION_TYPE'].replace(\n",
    "    'Waiters/barmen staff', 'WaitersBarmen')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['OCCUPATION_TYPE'], prefix='OCCT', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('OCCUPATION_TYPE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### CNT_FAM_MEMBERS\n",
    "How many family members does client have."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in CNT_FAM_MEMBERS: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in CNT_FAM_MEMBERS: {}\".format(\n",
    "    processed_app_train[processed_app_train.CNT_FAM_MEMBERS.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c07725ac8>"
      ]
     },
     "execution_count": 100,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['CNT_FAM_MEMBERS'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['CNT_FAM_MEMBERS'] = ((processed_app_train['CNT_FAM_MEMBERS'] -\n",
    "                                      processed_app_train['CNT_FAM_MEMBERS'].min())/\n",
    "                                     (processed_app_train['CNT_FAM_MEMBERS'].max() - \n",
    "                                      processed_app_train['CNT_FAM_MEMBERS'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REGION_RATING_CLIENT\n",
    "Our rating of the region where client lives (1,2,3). <br>\n",
    "Will be treated as categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 102,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in REGION_RATING_CLIENT: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in REGION_RATING_CLIENT: {}\".format(\n",
    "    processed_app_train[processed_app_train.REGION_RATING_CLIENT.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 103,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['REGION_RATING_CLIENT'], prefix='REGION', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('REGION_RATING_CLIENT',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REGION_RATING_CLIENT_W_CITY\n",
    "Our rating of the region where client lives with taking city into account (1,2,3). <br>\n",
    "Will be treated as categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 104,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in REGION_RATING_CLIENT_W_CITY: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in REGION_RATING_CLIENT_W_CITY: {}\".format(\n",
    "    processed_app_train[processed_app_train.REGION_RATING_CLIENT_W_CITY.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 105,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['REGION_RATING_CLIENT_W_CITY'], \n",
    "                                  prefix='REGIONCTY', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('REGION_RATING_CLIENT_W_CITY',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### WEEKDAY_APPR_PROCESS_START\n",
    "On which day of the week did the client apply for the loan. <br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 106,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in WEEKDAY_APPR_PROCESS_START: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in WEEKDAY_APPR_PROCESS_START: {}\".format(\n",
    "    processed_app_train[processed_app_train.WEEKDAY_APPR_PROCESS_START.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 107,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for WEEKDAY_APPR_PROCESS_START:  ['WEDNESDAY' 'MONDAY' 'THURSDAY' 'SUNDAY' 'SATURDAY' 'FRIDAY' 'TUESDAY']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for WEEKDAY_APPR_PROCESS_START: \", processed_app_train['WEEKDAY_APPR_PROCESS_START'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 108,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['WEEKDAY_APPR_PROCESS_START'], prefix='WD', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('WEEKDAY_APPR_PROCESS_START',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### HOUR_APPR_PROCESS_START\n",
    "Approximately at what hour did the client apply for the loan."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 109,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in HOUR_APPR_PROCESS_START: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in HOUR_APPR_PROCESS_START: {}\".format(\n",
    "    processed_app_train[processed_app_train.HOUR_APPR_PROCESS_START.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 110,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19db6d30>"
      ]
     },
     "execution_count": 110,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAYcAAAD8CAYAAACcjGjIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDMuMC4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvOIA7rQAAF15JREFUeJzt3X+M3PWd3/Hn6/h1VrgcEJIVNbSmjVWFxDqSrMBSqmpDTmDIHyZSaEEomByV0wiuiWS1caK2pCGo5A+SHihBchoXU3FxUH7UVnDqQxyrNFIgQEIxhEa4xA0GCsqZEJy0RObe/WM+LhN/x97Z2bVnvft8SKOZec/n853P97Mz+9rvj5lNVSFJUr8/GPcAJEkLj+EgSeowHCRJHYaDJKnDcJAkdRgOkqQOw0GS1GE4SJI6DAdJUseJ4x7AqM4888xasWLFSH1/85vf8KY3vWl+B3Qcch56nIce5+ENi3kuHn300V9W1VtnanfchsOKFSt45JFHRuo7PT3N1NTU/A7oOOQ89DgPPc7DGxbzXCT5X8O0c7eSJKnDcJAkdRgOkqQOw0GS1GE4SJI6DAdJUofhIEnqMBwkSR2GgySp47j9hLQ0X3Y99wrXbrx3xnZ7bvngMRiNtDC45SBJ6jAcJEkd7laShrRiiF1P4O4nLQ5uOUiSOgwHSVKH4SBJ6jAcJEkdhoMkqWPGcEjyh0l+lOS/J3kyyb9r9TuT/DzJY+1yfqsnyW1Jdid5PMl7+pa1LsnT7bKur/7eJLtan9uS5GisrCRpOMOcyvoacFFV7U9yEvCDJN9rj/3LqvrmIe0vBVa2y4XAHcCFSc4AbgQmgQIeTbK9ql5ubdYDDwI7gDXA95AkjcWMWw7Vs7/dPald6ghd1gJ3tX4PAqclOQu4BLivqva1QLgPWNMee3NV/bCqCrgLuHwO6yRJmqOhPgSX5ATgUeDtwJer6qEkHwduTvJvgfuBjVX1GrAceLav+95WO1J974D6oHGsp7eFwcTEBNPT08MMv2P//v0j911MnIeeiWWwYdWBeVve8Tqnvh7e4FwMGQ5V9TpwfpLTgO8keRfwaeB/AycDm4BPAZ8DBh0vqBHqg8axqT0Xk5OTNTU1NczwO6anpxm172LiPPTcfvc2bt01f18WsOfqqXlb1rHk6+ENzsUsz1aqql8B08Caqnqh7Tp6DfhPwAWt2V7gnL5uZwPPz1A/e0BdkjQmw5yt9Na2xUCSZcCfAv+jHSugnVl0OfBE67IduKadtbQaeKWqXgB2AhcnOT3J6cDFwM722KtJVrdlXQNsm9/VlCTNxjDb0mcBW9pxhz8A7qmq7yb56yRvpbdb6DHgn7f2O4DLgN3Ab4GPAlTVviQ3AQ+3dp+rqn3t9seBO4Fl9M5S8kwlSRqjGcOhqh4H3j2gftFh2hdw/WEe2wxsHlB/BHjXTGORJB0bfkJaktTh/3OQ5pn/90GLgVsOkqQOw0GS1GE4SJI6DAdJUofhIEnqMBwkSR2GgySpw3CQJHUYDpKkDsNBktRhOEiSOgwHSVKH4SBJ6jAcJEkdhoMkqcNwkCR1GA6SpI4Z/xNckj8Evg+c0tp/s6puTHIusBU4A/gx8JGq+l2SU4C7gPcCfwP806ra05b1aeA64HXgX1TVzlZfA/wFcALwH6vqlnldS2kB8j/GaSEbZsvhNeCiqvoT4HxgTZLVwBeAL1XVSuBler/0adcvV9XbgS+1diQ5D7gSeCewBvhKkhOSnAB8GbgUOA+4qrWVJI3JjOFQPfvb3ZPapYCLgG+2+hbg8nZ7bbtPe/wDSdLqW6vqtar6ObAbuKBddlfVM1X1O3pbI2vnvGaSpJHNuFsJoP11/yjwdnp/5f9P4FdVdaA12Qssb7eXA88CVNWBJK8Ab2n1B/sW29/n2UPqFx5mHOuB9QATExNMT08PM/yO/fv3j9x3MXEeeiaWwYZVB2ZuOCbH6mfk6+ENzsWQ4VBVrwPnJzkN+A7wjkHN2nUO89jh6oO2XmpAjaraBGwCmJycrKmpqSMP/DCmp6cZte9i4jz03H73Nm7dNdRbYSz2XD11TJ7H18MbnItZnq1UVb8CpoHVwGlJDr6jzgaeb7f3AucAtMf/GNjXXz+kz+HqkqQxmTEckry1bTGQZBnwp8BTwAPAh1uzdcC2dnt7u097/K+rqlr9yiSntDOdVgI/Ah4GViY5N8nJ9A5ab5+PlZMkjWaYbemzgC3tuMMfAPdU1XeT/BTYmuTzwE+Ar7X2XwP+c5Ld9LYYrgSoqieT3AP8FDgAXN92V5HkBmAnvVNZN1fVk/O2hpKkWZsxHKrqceDdA+rP0DvT6ND6/wWuOMyybgZuHlDfAewYYrySpGPAT0hLkjoMB0lSh+EgSeowHCRJHYaDJKnDcJAkdRgOkqQOw0GS1GE4SJI6Fu5XUUqaNf+7nOaL4aDjjr8ApaPP3UqSpA7DQZLUYThIkjoMB0lSh+EgSeowHCRJHYaDJKljxnBIck6SB5I8leTJJJ9o9c8meS7JY+1yWV+fTyfZneRnSS7pq69ptd1JNvbVz03yUJKnk3wjycnzvaKSpOENs+VwANhQVe8AVgPXJzmvPfalqjq/XXYAtMeuBN4JrAG+kuSEJCcAXwYuBc4DrupbzhfaslYCLwPXzdP6SZJGMGM4VNULVfXjdvtV4Clg+RG6rAW2VtVrVfVzYDdwQbvsrqpnqup3wFZgbZIAFwHfbP23AJePukKSpLmb1TGHJCuAdwMPtdINSR5PsjnJ6a22HHi2r9veVjtc/S3Ar6rqwCF1SdKYDP3dSklOBb4FfLKqfp3kDuAmoNr1rcCfARnQvRgcRHWE9oPGsB5YDzAxMcH09PSww/89+/fvH7nvYnK8zsOGVQdmbgRDr9vEsuGXOQ6z+RnNZW6O19fD0eBcDBkOSU6iFwx3V9W3Aarqxb7Hvwp8t93dC5zT1/1s4Pl2e1D9l8BpSU5sWw/97X9PVW0CNgFMTk7W1NTUMMPvmJ6eZtS+i8nxOg/XDvvFe1dPDdXu9ru3ceuuhfsdlMOuB8xtbo7X18PR4FwMd7ZSgK8BT1XVF/vqZ/U1+xDwRLu9HbgyySlJzgVWAj8CHgZWtjOTTqZ30Hp7VRXwAPDh1n8dsG1uqyVJmoth/lx6H/ARYFeSx1rtM/TONjqf3i6gPcDHAKrqyST3AD+ld6bT9VX1OkCSG4CdwAnA5qp6si3vU8DWJJ8HfkIvjCQdJYO+9nzDqgOdLQ+/9nzpmjEcquoHDD4usOMIfW4Gbh5Q3zGoX1U9Q+9sJknSAuAnpCVJHYaDJKnDcJAkdRgOkqQOw0GS1GE4SJI6DAdJUofhIEnqMBwkSR2GgySpw3CQJHUYDpKkDsNBktRhOEiSOgwHSVKH4SBJ6jAcJEkdhoMkqcNwkCR1zBgOSc5J8kCSp5I8meQTrX5GkvuSPN2uT2/1JLktye4kjyd5T9+y1rX2TydZ11d/b5Jdrc9tSQb9z2pJ0jEyzJbDAWBDVb0DWA1cn+Q8YCNwf1WtBO5v9wEuBVa2y3rgDuiFCXAjcCFwAXDjwUBpbdb39Vsz91WTJI1qxnCoqheq6sft9qvAU8ByYC2wpTXbAlzebq8F7qqeB4HTkpwFXALcV1X7qupl4D5gTXvszVX1w6oq4K6+ZUmSxuDE2TROsgJ4N/AQMFFVL0AvQJK8rTVbDjzb121vqx2pvndAXdKYrdh471Dt9tzywaM8Eh1rQ4dDklOBbwGfrKpfH+GwwKAHaoT6oDGsp7f7iYmJCaanp2cY9WD79+8fue9icrzOw4ZVB4ZqN+y6TSwbfpnjMJuf0VzWYy7zcDy+jo7keH1vzKehwiHJSfSC4e6q+nYrv5jkrLbVcBbwUqvvBc7p63428HyrTx1Sn271swe076iqTcAmgMnJyZqamhrUbEbT09OM2ncxOV7n4dph/5q9emqodrffvY1bd81qI/qYGnY9YPi5GWTDqgMjz8Nsxng8OF7fG/NpmLOVAnwNeKqqvtj30Hbg4BlH64BtffVr2llLq4FX2u6nncDFSU5vB6IvBna2x15Nsro91zV9y5IkjcEwfya8D/gIsCvJY632GeAW4J4k1wG/AK5oj+0ALgN2A78FPgpQVfuS3AQ83Np9rqr2tdsfB+4ElgHfaxdJ0pjMGA5V9QMGHxcA+MCA9gVcf5hlbQY2D6g/ArxrprFIko4NPyEtSeowHCRJHYaDJKnDcJAkdRgOkqQOw0GS1GE4SJI6DAdJUofhIEnqMBwkSR2GgySpw3CQJHUYDpKkDsNBktRhOEiSOgwHSVKH4SBJ6jAcJEkdhoMkqWPGcEiyOclLSZ7oq302yXNJHmuXy/oe+3SS3Ul+luSSvvqaVtudZGNf/dwkDyV5Osk3kpw8nysoSZq9YbYc7gTWDKh/qarOb5cdAEnOA64E3tn6fCXJCUlOAL4MXAqcB1zV2gJ8oS1rJfAycN1cVkiSNHczhkNVfR/YN+Ty1gJbq+q1qvo5sBu4oF12V9UzVfU7YCuwNkmAi4Bvtv5bgMtnuQ6SpHk2l2MONyR5vO12Or3VlgPP9rXZ22qHq78F+FVVHTikLkkaoxNH7HcHcBNQ7fpW4M+ADGhbDA6hOkL7gZKsB9YDTExMMD09PatBH7R///6R+y4mC2kedj33ytBtN6wart2w6zaxDDasOjBzwzGZzc9oLusxl3lYKK+j+bKQ3hvjMlI4VNWLB28n+Srw3XZ3L3BOX9Ozgefb7UH1XwKnJTmxbT30tx/0vJuATQCTk5M1NTU1yvCZnp5m1L6LyUKah2s33jvvy9xz9dRQ7W6/exu37hr176Sjb9j1gLnN44ZVB0aeh2HHuGLI8e255YMjjWO+LKT3xriMtFspyVl9dz8EHDyTaTtwZZJTkpwLrAR+BDwMrGxnJp1M76D19qoq4AHgw63/OmDbKGOSJM2fGf9MSPJ1YAo4M8le4EZgKsn59HYB7QE+BlBVTya5B/gpcAC4vqpeb8u5AdgJnABsrqon21N8Ctia5PPAT4CvzdvaSZJGMmM4VNVVA8qH/QVeVTcDNw+o7wB2DKg/Q+9sJknSAuEnpCVJHYaDJKnDcJAkdRgOkqQOw0GS1GE4SJI6DAdJUofhIEnqMBwkSR2GgySpw3CQJHUYDpKkDsNBktRhOEiSOgwHSVKH4SBJ6jAcJEkdhoMkqcNwkCR1zBgOSTYneSnJE321M5Lcl+Tpdn16qyfJbUl2J3k8yXv6+qxr7Z9Osq6v/t4ku1qf25JkvldSkjQ7w2w53AmsOaS2Ebi/qlYC97f7AJcCK9tlPXAH9MIEuBG4ELgAuPFgoLQ26/v6HfpckqRjbMZwqKrvA/sOKa8FtrTbW4DL++p3Vc+DwGlJzgIuAe6rqn1V9TJwH7CmPfbmqvphVRVwV9+yJEljcuKI/Saq6gWAqnohydtafTnwbF+7va12pPreAfWBkqynt5XBxMQE09PTIw1+//79I/ddTBbSPGxYdWDelznsuk0sOzrPP19m8zOay3rMZR6GHeOwyx/363IhvTfGZdRwOJxBxwtqhPpAVbUJ2AQwOTlZU1NTIwyx98Ibte9ispDm4dqN9877MvdcPTVUu9vv3satu+b7rTB/hl0PmNs8blh1YOR5GHaMw45vNut8NCyk98a4jHq20ottlxDt+qVW3wuc09fubOD5GepnD6hLksZo1HDYDhw842gdsK2vfk07a2k18Erb/bQTuDjJ6e1A9MXAzvbYq0lWt7OUrulbliRpTGbchkzydWAKODPJXnpnHd0C3JPkOuAXwBWt+Q7gMmA38FvgowBVtS/JTcDDrd3nqurgQe6P0zsjahnwvXaRJI3RjOFQVVcd5qEPDGhbwPWHWc5mYPOA+iPAu2YahyTp2Fm4R+G0KKw4CgeatfgN+7rZc8sHj/JIli6/PkOS1GE4SJI6DAdJUofhIEnqMBwkSR2GgySpw3CQJHUYDpKkDsNBktRhOEiSOgwHSVKH4SBJ6jAcJEkdhoMkqcNwkCR1GA6SpA7DQZLUMaf/BJdkD/Aq8DpwoKomk5wBfANYAewB/klVvZwkwF/Q+x/TvwWuraoft+WsA/51W+znq2rLXMal0fkfuCTB/Gw5vL+qzq+qyXZ/I3B/Va0E7m/3AS4FVrbLeuAOgBYmNwIXAhcANyY5fR7GJUka0dHYrbQWOPiX/xbg8r76XdXzIHBakrOAS4D7qmpfVb0M3AesOQrjkiQNaa7hUMBfJXk0yfpWm6iqFwDa9dtafTnwbF/fva12uLokaUxSVaN3Tv5OVT2f5G30/uL/c2B7VZ3W1+blqjo9yb3Av6+qH7T6/cC/Ai4CTqmqz7f6vwF+W1W3Dni+9fR2STExMfHerVu3jjTu/fv3c+qpp47UdzEZNA+7nntlqL6rlv/xUO2GXd7RMOwYX9r3Ci/+n6M8mDkYdj1gbvM9sYyR52Fcr4fZzM1sLObfEe9///sf7TsMcFhzOiBdVc+365eSfIfeMYMXk5xVVS+03UYvteZ7gXP6up8NPN/qU4fUpw/zfJuATQCTk5M1NTU1qNmMpqenGbXvYjJoHq4d9oD01VMztpnN8o6GYcd4+93buHXXnN4KR9Ww6wFzm+8Nqw6MPA/jej3MZm5mw98Rc9itlORNSf7o4G3gYuAJYDuwrjVbB2xrt7cD16RnNfBK2+20E7g4yentQPTFrSZJGpO5/Lk0AXynd4YqJwJ/WVX/NcnDwD1JrgN+AVzR2u+gdxrrbnqnsn4UoKr2JbkJeLi1+1xV7ZvDuCRJczRyOFTVM8CfDKj/DfCBAfUCrj/MsjYDm0cdiyRpfvkJaUlSh+EgSepYuKdoSNIMhv26F/ArX2bLcFgiBr2JNqw6MNZTTSUtXO5WkiR1GA6SpA7DQZLUYThIkjoMB0lSh+EgSeowHCRJHX7O4Tg3mw8BSdKw3HKQJHUYDpKkDsNBktRhOEiSOjwgfQz5DZLSwrdi471DfSnlYn+PuuUgSeowHCRJHQsmHJKsSfKzJLuTbBz3eCRpKVsQ4ZDkBODLwKXAecBVSc4b76gkaelaKAekLwB2V9UzAEm2AmuBn451VEPyU8rS0jPs+/54PXC9UMJhOfBs3/29wIVjGsv/5y99SXN1vIZIqmrcYyDJFcAlVfXP2v2PABdU1Z8f0m49sL7d/YfAz0Z8yjOBX47YdzFxHnqchx7n4Q2LeS7+XlW9daZGC2XLYS9wTt/9s4HnD21UVZuATXN9siSPVNXkXJdzvHMeepyHHufhDc7FAjkgDTwMrExybpKTgSuB7WMekyQtWQtiy6GqDiS5AdgJnABsrqonxzwsSVqyFkQ4AFTVDmDHMXq6Oe+aWiSchx7nocd5eMOSn4sFcUBakrSwLJRjDpKkBWRJhYNf0fGGJHuS7EryWJJHxj2eYyXJ5iQvJXmir3ZGkvuSPN2uTx/nGI+Fw8zDZ5M8114TjyW5bJxjPBaSnJPkgSRPJXkyySdafcm9Jg61ZMLBr+gY6P1Vdf4SO2XvTmDNIbWNwP1VtRK4v91f7O6kOw8AX2qvifPbccDF7gCwoareAawGrm+/F5bia+L3LJlwoO8rOqrqd8DBr+jQElJV3wf2HVJeC2xpt7cAlx/TQY3BYeZhyamqF6rqx+32q8BT9L6xYcm9Jg61lMJh0Fd0LB/TWBaCAv4qyaPtk+dL2URVvQC9XxbA28Y8nnG6IcnjbbfTktqVkmQF8G7gIXxNLKlwyIDaUj5V631V9R56u9muT/KPxz0gjd0dwD8AzgdeAG4d73COnSSnAt8CPllVvx73eBaCpRQOQ31Fx1JRVc+365eA79Db7bZUvZjkLIB2/dKYxzMWVfViVb1eVX8LfJUl8ppIchK9YLi7qr7dykv+NbGUwsGv6GiSvCnJHx28DVwMPHHkXovadmBdu70O2DbGsYzNwV+GzYdYAq+JJAG+BjxVVV/se2jJvyaW1Ifg2ql5/4E3vqLj5jEPaSyS/H16WwvQ+5T8Xy6VuUjydWCK3rduvgjcCPwX4B7g7wK/AK6oqkV9sPYw8zBFb5dSAXuAjx3c775YJflHwH8DdgF/28qfoXfcYUm9Jg61pMJBkjScpbRbSZI0JMNBktRhOEiSOgwHSVKH4SBJ6jAcJEkdhoMkqcNwkCR1/D822OIqZzLQawAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['HOUR_APPR_PROCESS_START'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 111,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['HOUR_APPR_PROCESS_START'] = ((processed_app_train['HOUR_APPR_PROCESS_START'] -\n",
    "                                      processed_app_train['HOUR_APPR_PROCESS_START'].min())/\n",
    "                                     (processed_app_train['HOUR_APPR_PROCESS_START'].max() - \n",
    "                                      processed_app_train['HOUR_APPR_PROCESS_START'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REG_REGION_NOT_LIVE_REGION\n",
    "Flag if client's permanent address does not match contact address (1=different, 0=same, at region level)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 112,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in REG_REGION_NOT_LIVE_REGION: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in REG_REGION_NOT_LIVE_REGION: {}\".format(\n",
    "    processed_app_train[processed_app_train.REG_REGION_NOT_LIVE_REGION.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REG_REGION_NOT_WORK_REGION\n",
    "Flag if client's permanent address does not match work address (1=different, 0=same, at region level)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 113,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in REG_REGION_NOT_WORK_REGION: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in REG_REGION_NOT_WORK_REGION: {}\".format(\n",
    "    processed_app_train[processed_app_train.REG_REGION_NOT_WORK_REGION.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LIVE_REGION_NOT_WORK_REGION\n",
    "Flag if client's contact address does not match work address (1=different, 0=same, at region level)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 114,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in LIVE_REGION_NOT_WORK_REGION: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in LIVE_REGION_NOT_WORK_REGION: {}\".format(\n",
    "    processed_app_train[processed_app_train.LIVE_REGION_NOT_WORK_REGION.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REG_CITY_NOT_LIVE_CITY\n",
    "Flag if client's permanent address does not match contact address (1=different, 0=same, at city level)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 115,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in REG_CITY_NOT_LIVE_CITY: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in REG_CITY_NOT_LIVE_CITY: {}\".format(\n",
    "    processed_app_train[processed_app_train.REG_CITY_NOT_LIVE_CITY.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### REG_CITY_NOT_WORK_CITY\n",
    "Flag if client's permanent address does not match work address (1=different, 0=same, at city level)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 116,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in REG_CITY_NOT_WORK_CITY: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in REG_CITY_NOT_WORK_CITY: {}\".format(\n",
    "    processed_app_train[processed_app_train.REG_CITY_NOT_WORK_CITY.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### LIVE_CITY_NOT_WORK_CITY\n",
    "Flag if client's contact address does not match work address (1=different, 0=same, at city level)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 117,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in LIVE_CITY_NOT_WORK_CITY: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in LIVE_CITY_NOT_WORK_CITY: {}\".format(\n",
    "    processed_app_train[processed_app_train.LIVE_CITY_NOT_WORK_CITY.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### ORGANIZATION_TYPE\n",
    "Type of organization where client works. <br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 118,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in ORGANIZATION_TYPE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in ORGANIZATION_TYPE: {}\".format(\n",
    "    processed_app_train[processed_app_train.ORGANIZATION_TYPE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 119,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for ORGANIZATION_TYPE:  ['Business Entity Type 3' 'School' 'Government' 'Religion' 'Other' 'XNA'\n",
      " 'Electricity' 'Medicine' 'Business Entity Type 2' 'Self-employed'\n",
      " 'Transport: type 2' 'Construction' 'Housing' 'Kindergarten'\n",
      " 'Trade: type 7' 'Industry: type 11' 'Military' 'Services'\n",
      " 'Security Ministries' 'Transport: type 4' 'Industry: type 1' 'Emergency'\n",
      " 'Security' 'Trade: type 2' 'University' 'Transport: type 3' 'Police'\n",
      " 'Business Entity Type 1' 'Postal' 'Industry: type 4' 'Agriculture'\n",
      " 'Restaurant' 'Culture' 'Hotel' 'Industry: type 7' 'Trade: type 3'\n",
      " 'Industry: type 3' 'Bank' 'Industry: type 9' 'Insurance' 'Trade: type 6'\n",
      " 'Industry: type 2' 'Transport: type 1' 'Industry: type 12' 'Mobile'\n",
      " 'Trade: type 1' 'Industry: type 5' 'Industry: type 10' 'Legal Services'\n",
      " 'Advertising' 'Trade: type 5' 'Cleaning' 'Industry: type 13'\n",
      " 'Trade: type 4' 'Telecom' 'Industry: type 8' 'Realtor' 'Industry: type 6']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for ORGANIZATION_TYPE: \", processed_app_train['ORGANIZATION_TYPE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 120,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['ORGANIZATION_TYPE'], prefix='ORGT', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('ORGANIZATION_TYPE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EXT_SOURCE_1\n",
    "Normalized score from external data source."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 121,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in EXT_SOURCE_1: 171450\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in EXT_SOURCE_1: {}\".format(\n",
    "    processed_app_train[processed_app_train.EXT_SOURCE_1.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 122,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fill the missing values with the mean value\n",
    "processed_app_train['EXT_SOURCE_1'].fillna(value=processed_app_train['EXT_SOURCE_1'].mean(), inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EXT_SOURCE_2\n",
    "Normalized score from external data source."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 123,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in EXT_SOURCE_2: 649\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in EXT_SOURCE_2: {}\".format(\n",
    "    processed_app_train[processed_app_train.EXT_SOURCE_2.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 124,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Few rows with nulls -> delete rows\n",
    "processed_app_train = processed_app_train[processed_app_train.EXT_SOURCE_2.notnull()]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EXT_SOURCE_3\n",
    "Normalized score from external data source."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 125,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in EXT_SOURCE_3: 60149\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in EXT_SOURCE_3: {}\".format(\n",
    "    processed_app_train[processed_app_train.EXT_SOURCE_3.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 126,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Fill the missing values with the mean value\n",
    "processed_app_train['EXT_SOURCE_3'].fillna(value=processed_app_train['EXT_SOURCE_3'].mean(), inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 127,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "303046"
      ]
     },
     "execution_count": 127,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train.shape[0]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FONDKAPREMONT_MODE\n",
    "Normalized information about building where the client lives, What is average (_AVG suffix), modus (_MODE suffix), median (_MEDI suffix) apartment size, common area, living area, age of building, number of elevators, number of entrances, state of the building, number of floor.<br>\n",
    "Categorial."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 128,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in FONDKAPREMONT_MODE: 207311\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in FONDKAPREMONT_MODE: {}\".format(\n",
    "    processed_app_train[processed_app_train.FONDKAPREMONT_MODE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 129,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for FONDKAPREMONT_MODE:  ['reg oper account' nan 'org spec account' 'reg oper spec account'\n",
      " 'not specified']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for FONDKAPREMONT_MODE: \", processed_app_train['FONDKAPREMONT_MODE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 130,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['FONDKAPREMONT_MODE'].fillna(value='Unspec', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 131,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['FONDKAPREMONT_MODE'], prefix='FONDKA', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('FONDKAPREMONT_MODE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### HOUSETYPE_MODE\n",
    "Normalized information about building where the client lives, What is average (_AVG suffix), modus (_MODE suffix), median (_MEDI suffix) apartment size, common area, living area, age of building, number of elevators, number of entrances, state of the building, number of floor.<br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 132,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in HOUSETYPE_MODE: 152135\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in HOUSETYPE_MODE: {}\".format(\n",
    "    processed_app_train[processed_app_train.HOUSETYPE_MODE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 133,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for HOUSETYPE_MODE:  ['block of flats' nan 'terraced house' 'specific housing']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for HOUSETYPE_MODE: \", processed_app_train['HOUSETYPE_MODE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 134,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['HOUSETYPE_MODE'].fillna(value='Unspec', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 135,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['HOUSETYPE_MODE'], prefix='HTM', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('HOUSETYPE_MODE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### WALLSMATERIAL_MODE\n",
    "Normalized information about building where the client lives, What is average (_AVG suffix), modus (_MODE suffix), median (_MEDI suffix) apartment size, common area, living area, age of building, number of elevators, number of entrances, state of the building, number of floor. <br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 136,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in WALLSMATERIAL_MODE: 154156\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in WALLSMATERIAL_MODE: {}\".format(\n",
    "    processed_app_train[processed_app_train.WALLSMATERIAL_MODE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 137,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['WALLSMATERIAL_MODE'].fillna(value='Unspec', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 138,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['WALLSMATERIAL_MODE'], prefix='WMOD', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('WALLSMATERIAL_MODE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### EMERGENCYSTATE_MODE\n",
    "Normalized information about building where the client lives, What is average (_AVG suffix), modus (_MODE suffix), median (_MEDI suffix) apartment size, common area, living area, age of building, number of elevators, number of entrances, state of the building, number of floor.<br>\n",
    "Categorical."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 139,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in EMERGENCYSTATE_MODE: 143699\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in EMERGENCYSTATE_MODE: {}\".format(\n",
    "    processed_app_train[processed_app_train.EMERGENCYSTATE_MODE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 140,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Values for EMERGENCYSTATE_MODE:  ['No' nan 'Yes']\n"
     ]
    }
   ],
   "source": [
    "# List values\n",
    "print(\"Values for EMERGENCYSTATE_MODE: \", processed_app_train['EMERGENCYSTATE_MODE'].unique())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 141,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Preparing data for One Hot Encoding\n",
    "processed_app_train['EMERGENCYSTATE_MODE'].fillna(value='Unspec', inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 142,
   "metadata": {},
   "outputs": [],
   "source": [
    "# One hot encoding\n",
    "one_hot_encoding = pd.get_dummies(processed_app_train['EMERGENCYSTATE_MODE'], prefix='EMGMODE', prefix_sep='_')\n",
    "processed_app_train = processed_app_train.drop('EMERGENCYSTATE_MODE',axis = 1)\n",
    "processed_app_train = pd.concat([processed_app_train, one_hot_encoding], axis=1)\n",
    "processed_app_train.columns = processed_app_train.columns.str.strip().str.upper().str.replace(' ', '_')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OBS_30_CNT_SOCIAL_CIRCLE\n",
    "How many observation of client's social surroundings with observable 30 DPD (days past due) default."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 143,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in OBS_30_CNT_SOCIAL_CIRCLE: 992\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in OBS_30_CNT_SOCIAL_CIRCLE: {}\".format(\n",
    "    processed_app_train[processed_app_train.OBS_30_CNT_SOCIAL_CIRCLE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 144,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Few rows with nulls -> delete rows\n",
    "processed_app_train = processed_app_train[processed_app_train.OBS_30_CNT_SOCIAL_CIRCLE.notnull()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 145,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c1ab39e80>"
      ]
     },
     "execution_count": 145,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 146,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 12.271381467680962\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 147,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19b13080>"
      ]
     },
     "execution_count": 147,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'].replace(0,0.00001).apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 148,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'] = processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'].replace(0,0.00001)\n",
    "processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'] = np.log(processed_app_train.OBS_30_CNT_SOCIAL_CIRCLE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 149,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'] = ((processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'] -\n",
    "                                      processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'].min())/\n",
    "                                     (processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'].max() - \n",
    "                                      processed_app_train['OBS_30_CNT_SOCIAL_CIRCLE'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DEF_30_CNT_SOCIAL_CIRCLE\n",
    "How many observation of client's social surroundings defaulted on 30 DPD (days past due)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 150,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in DEF_30_CNT_SOCIAL_CIRCLE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in DEF_30_CNT_SOCIAL_CIRCLE: {}\".format(\n",
    "    processed_app_train[processed_app_train.DEF_30_CNT_SOCIAL_CIRCLE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 151,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19bbcf28>"
      ]
     },
     "execution_count": 151,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 152,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 5.201264871756483\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 153,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19c58080>"
      ]
     },
     "execution_count": 153,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'].replace(0,0.00001).apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 154,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'] = processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'].replace(0,0.00001)\n",
    "processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'] = np.log(processed_app_train.DEF_30_CNT_SOCIAL_CIRCLE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 155,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'] = ((processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'] -\n",
    "                                      processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'].min())/\n",
    "                                     (processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'].max() - \n",
    "                                      processed_app_train['DEF_30_CNT_SOCIAL_CIRCLE'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### OBS_60_CNT_SOCIAL_CIRCLE\n",
    "How many observation of client's social surroundings with observable 60 DPD (days past due) default."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 156,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in OBS_60_CNT_SOCIAL_CIRCLE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in OBS_60_CNT_SOCIAL_CIRCLE: {}\".format(\n",
    "    processed_app_train[processed_app_train.OBS_60_CNT_SOCIAL_CIRCLE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 157,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19d0a550>"
      ]
     },
     "execution_count": 157,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 158,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 12.201071954709487\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 159,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19e94d68>"
      ]
     },
     "execution_count": 159,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'].replace(0,0.00001).apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 160,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'] = processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'].replace(0,0.00001)\n",
    "processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'] = np.log(processed_app_train.OBS_60_CNT_SOCIAL_CIRCLE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 161,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'] = ((processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'] -\n",
    "                                      processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'].min())/\n",
    "                                     (processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'].max() - \n",
    "                                      processed_app_train['OBS_60_CNT_SOCIAL_CIRCLE'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DEF_60_CNT_SOCIAL_CIRCLE\n",
    "How many observation of client's social surroundings defaulted on 60 (days past due) DPD."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 162,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in DEF_60_CNT_SOCIAL_CIRCLE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in DEF_60_CNT_SOCIAL_CIRCLE: {}\".format(\n",
    "    processed_app_train[processed_app_train.DEF_60_CNT_SOCIAL_CIRCLE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 163,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19f35da0>"
      ]
     },
     "execution_count": 163,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 164,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Skewness: 5.291578281898555\n"
     ]
    }
   ],
   "source": [
    "# Data seems to be skewed\n",
    "print(\"Skewness: {}\".format(processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'].skew()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 165,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c19f8e978>"
      ]
     },
     "execution_count": 165,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Data skewed: try to apply log\n",
    "processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'].replace(0,0.00001).apply(np.log).hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 166,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Apply log\n",
    "processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'] = processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'].replace(0,0.00001)\n",
    "processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'] = np.log(processed_app_train.DEF_60_CNT_SOCIAL_CIRCLE)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 167,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'] = ((processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'] -\n",
    "                                      processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'].min())/\n",
    "                                     (processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'].max() - \n",
    "                                      processed_app_train['DEF_60_CNT_SOCIAL_CIRCLE'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### DAYS_LAST_PHONE_CHANGE\n",
    "How many days before application did client change phone."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 168,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in DAYS_LAST_PHONE_CHANGE: 0\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in DAYS_LAST_PHONE_CHANGE: {}\".format(\n",
    "    processed_app_train[processed_app_train.DAYS_LAST_PHONE_CHANGE.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 169,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x23c1a338a20>"
      ]
     },
     "execution_count": 169,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Check the data distribution\n",
    "processed_app_train['DAYS_LAST_PHONE_CHANGE'].hist(bins=30)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 170,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['DAYS_LAST_PHONE_CHANGE'] = ((processed_app_train['DAYS_LAST_PHONE_CHANGE'] -\n",
    "                                      processed_app_train['DAYS_LAST_PHONE_CHANGE'].min())/\n",
    "                                     (processed_app_train['DAYS_LAST_PHONE_CHANGE'].max() - \n",
    "                                      processed_app_train['DAYS_LAST_PHONE_CHANGE'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### FLAG_DOCUMENT_2\n",
    "Did client provide document 2."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 171,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    302054.000000\n",
       "mean          0.000043\n",
       "std           0.006560\n",
       "min           0.000000\n",
       "25%           0.000000\n",
       "50%           0.000000\n",
       "75%           0.000000\n",
       "max           1.000000\n",
       "Name: FLAG_DOCUMENT_2, dtype: float64"
      ]
     },
     "execution_count": 171,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train['FLAG_DOCUMENT_2'].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "There is no need to change the values od FLAG_DOCUMENT_2 to 21. Values between 0 and 1 with no missing values.\n",
    "\n",
    "### AMT_REQ_CREDIT_BUREAU_HOUR\n",
    "Number of enquiries to Credit Bureau about the client one hour before application."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 172,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Rows with missing values in AMT_REQ_CREDIT_BUREAU_HOUR: 40800\n"
     ]
    }
   ],
   "source": [
    "# Checking for missing values\n",
    "print(\"Rows with missing values in AMT_REQ_CREDIT_BUREAU_HOUR: {}\".format(\n",
    "    processed_app_train[processed_app_train.AMT_REQ_CREDIT_BUREAU_HOUR.isnull()].shape[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 173,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    261254.000000\n",
       "mean          0.006431\n",
       "std           0.084041\n",
       "min           0.000000\n",
       "25%           0.000000\n",
       "50%           0.000000\n",
       "75%           0.000000\n",
       "max           4.000000\n",
       "Name: AMT_REQ_CREDIT_BUREAU_HOUR, dtype: float64"
      ]
     },
     "execution_count": 173,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 174,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'].fillna(value=processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'].\n",
    "                                                        mean(),inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 175,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'] = ((processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'] -\n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'].min())/\n",
    "                                     (processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'].max() - \n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_HOUR'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_REQ_CREDIT_BUREAU_DAY\n",
    "Number of enquiries to Credit Bureau about the client one day before application (excluding one hour before application)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 176,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    261254.000000\n",
       "mean          0.007051\n",
       "std           0.111140\n",
       "min           0.000000\n",
       "25%           0.000000\n",
       "50%           0.000000\n",
       "75%           0.000000\n",
       "max           9.000000\n",
       "Name: AMT_REQ_CREDIT_BUREAU_DAY, dtype: float64"
      ]
     },
     "execution_count": 176,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 177,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'].fillna(value=processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'].\n",
    "                                                        mean(),inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 178,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'] = ((processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'] -\n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'].min())/\n",
    "                                     (processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'].max() - \n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_DAY'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_REQ_CREDIT_BUREAU_WEEK\n",
    "Number of enquiries to Credit Bureau about the client one week before application (excluding one day before application)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 179,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    261254.000000\n",
       "mean          0.034361\n",
       "std           0.203871\n",
       "min           0.000000\n",
       "25%           0.000000\n",
       "50%           0.000000\n",
       "75%           0.000000\n",
       "max           8.000000\n",
       "Name: AMT_REQ_CREDIT_BUREAU_WEEK, dtype: float64"
      ]
     },
     "execution_count": 179,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 180,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'].fillna(value=processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'].\n",
    "                                                        mean(),inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 181,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'] = ((processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'] -\n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'].min())/\n",
    "                                     (processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'].max() - \n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_WEEK'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_REQ_CREDIT_BUREAU_MON\n",
    "Number of enquiries to Credit Bureau about the client one month before application (excluding one week before application)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 182,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    261254.000000\n",
       "mean          0.267169\n",
       "std           0.915371\n",
       "min           0.000000\n",
       "25%           0.000000\n",
       "50%           0.000000\n",
       "75%           0.000000\n",
       "max          27.000000\n",
       "Name: AMT_REQ_CREDIT_BUREAU_MON, dtype: float64"
      ]
     },
     "execution_count": 182,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 183,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'].fillna(value=processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'].\n",
    "                                                        mean(),inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 184,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'] = ((processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'] -\n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'].min())/\n",
    "                                     (processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'].max() - \n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_MON'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_REQ_CREDIT_BUREAU_QRT\n",
    "Number of enquiries to Credit Bureau about the client 3 month before application (excluding one month before application)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 185,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    261254.000000\n",
       "mean          0.266511\n",
       "std           0.797870\n",
       "min           0.000000\n",
       "25%           0.000000\n",
       "50%           0.000000\n",
       "75%           0.000000\n",
       "max         261.000000\n",
       "Name: AMT_REQ_CREDIT_BUREAU_QRT, dtype: float64"
      ]
     },
     "execution_count": 185,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 186,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'].fillna(value=processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'].\n",
    "                                                        mean(),inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 187,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'] = ((processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'] -\n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'].min())/\n",
    "                                     (processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'].max() - \n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_QRT'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### AMT_REQ_CREDIT_BUREAU_YEAR\n",
    "Number of enquiries to Credit Bureau about the client one day year (excluding last 3 months before application)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 188,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    261254.000000\n",
       "mean          1.906669\n",
       "std           1.872718\n",
       "min           0.000000\n",
       "25%           0.000000\n",
       "50%           1.000000\n",
       "75%           3.000000\n",
       "max          25.000000\n",
       "Name: AMT_REQ_CREDIT_BUREAU_YEAR, dtype: float64"
      ]
     },
     "execution_count": 188,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'].describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 189,
   "metadata": {},
   "outputs": [],
   "source": [
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'].fillna(value=processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'].\n",
    "                                                         mean(),inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 190,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'] = ((processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'] -\n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'].min())/\n",
    "                                     (processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'].max() - \n",
    "                                      processed_app_train['AMT_REQ_CREDIT_BUREAU_YEAR'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "\n",
    "For the remaining tables I will try to group and extract meaningful statistics consolidate it as new features of the training table.\n",
    "\n",
    "## TABLE bureau_balance.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(27299925, 3)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau_balance = pd.read_csv(\"data/bureau_balance.csv\")\n",
    "bureau_balance.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## STATUS\n",
    "Status of Credit Bureau loan during the month (active, closed, DPD0-30,… [C means closed, X means status unknown, 0 means no DPD, 1 means maximal did during month between 1-30, 2 means DPD 31-60,… 5 means DPD 120+ or sold or written off ] )."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 192,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "bureau_balance['STATUS'] = bureau_balance['STATUS'].replace('X', np.NaN)\n",
    "bureau_balance['STATUS'] = bureau_balance['STATUS'].replace('C', '0')\n",
    "bureau_balance['STATUS'] = pd.to_numeric(bureau_balance['STATUS'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 193,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_BUREAU</th>\n",
       "      <th>STATUS_MAX</th>\n",
       "      <th>STATUS_MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>5001709</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5001710</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>5001711</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>5001712</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5001713</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_BUREAU  STATUS_MAX  STATUS_MEAN\n",
       "0       5001709         0.0          0.0\n",
       "1       5001710         0.0          0.0\n",
       "2       5001711         0.0          0.0\n",
       "3       5001712         0.0          0.0\n",
       "4       5001713         NaN          NaN"
      ]
     },
     "execution_count": 193,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_bureau_balance = bureau_balance.groupby('SK_ID_BUREAU').agg({'STATUS':['max','mean']})\n",
    "agg_bureau_balance.columns = agg_bureau_balance.columns.map('_'.join).str.strip('_')\n",
    "agg_bureau_balance.columns = map(str.upper, agg_bureau_balance.columns)\n",
    "agg_bureau_balance.reset_index(inplace=True)\n",
    "agg_bureau_balance.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 194,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "agg_bureau_balance['STATUS_MAX'] = ((agg_bureau_balance['STATUS_MAX'] -\n",
    "                                      agg_bureau_balance['STATUS_MAX'].min())/\n",
    "                                     (agg_bureau_balance['STATUS_MAX'].max() - \n",
    "                                      agg_bureau_balance['STATUS_MAX'].min()))\n",
    "\n",
    "agg_bureau_balance['STATUS_MEAN'] = ((agg_bureau_balance['STATUS_MEAN'] -\n",
    "                                      agg_bureau_balance['STATUS_MEAN'].min())/\n",
    "                                     (agg_bureau_balance['STATUS_MEAN'].max() - \n",
    "                                      agg_bureau_balance['STATUS_MEAN'].min()))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TABLE bureau.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1716428, 17)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "bureau = pd.read_csv(\"data/bureau.csv\")\n",
    "bureau.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 196,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_bureau = pd.merge(bureau, agg_bureau_balance, how='left', on='SK_ID_BUREAU')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 197,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>CREDIT_DAY_OVERDUE_MAX</th>\n",
       "      <th>CREDIT_DAY_OVERDUE_MEAN</th>\n",
       "      <th>AMT_CREDIT_MAX_OVERDUE_MAX</th>\n",
       "      <th>CNT_CREDIT_PROLONG_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_MEAN</th>\n",
       "      <th>AMT_CREDIT_SUM_DEBT_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_DEBT_MEAN</th>\n",
       "      <th>AMT_CREDIT_SUM_LIMIT_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_LIMIT_MEAN</th>\n",
       "      <th>AMT_CREDIT_SUM_OVERDUE_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_OVERDUE_MEAN</th>\n",
       "      <th>AMT_ANNUITY_MAX</th>\n",
       "      <th>AMT_ANNUITY_MEAN</th>\n",
       "      <th>STATUS_MAX_MAX</th>\n",
       "      <th>STATUS_MEAN_MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "      <td>378000.0</td>\n",
       "      <td>207623.571429</td>\n",
       "      <td>373239.0</td>\n",
       "      <td>85240.928571</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10822.5</td>\n",
       "      <td>3545.357143</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.002198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>5043.645</td>\n",
       "      <td>0</td>\n",
       "      <td>450000.0</td>\n",
       "      <td>108131.945625</td>\n",
       "      <td>245781.0</td>\n",
       "      <td>49156.200000</td>\n",
       "      <td>31988.565</td>\n",
       "      <td>7997.14125</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.059844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0</td>\n",
       "      <td>810000.0</td>\n",
       "      <td>254350.125000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>810000.000</td>\n",
       "      <td>202500.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0</td>\n",
       "      <td>94537.8</td>\n",
       "      <td>94518.900000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0</td>\n",
       "      <td>568800.0</td>\n",
       "      <td>219042.000000</td>\n",
       "      <td>543087.0</td>\n",
       "      <td>189469.500000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.00000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>4261.5</td>\n",
       "      <td>1420.500000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  CREDIT_DAY_OVERDUE_MAX  CREDIT_DAY_OVERDUE_MEAN  \\\n",
       "0      100001                       0                      0.0   \n",
       "1      100002                       0                      0.0   \n",
       "2      100003                       0                      0.0   \n",
       "3      100004                       0                      0.0   \n",
       "4      100005                       0                      0.0   \n",
       "\n",
       "   AMT_CREDIT_MAX_OVERDUE_MAX  CNT_CREDIT_PROLONG_MAX  AMT_CREDIT_SUM_MAX  \\\n",
       "0                         NaN                       0            378000.0   \n",
       "1                    5043.645                       0            450000.0   \n",
       "2                       0.000                       0            810000.0   \n",
       "3                       0.000                       0             94537.8   \n",
       "4                       0.000                       0            568800.0   \n",
       "\n",
       "   AMT_CREDIT_SUM_MEAN  AMT_CREDIT_SUM_DEBT_MAX  AMT_CREDIT_SUM_DEBT_MEAN  \\\n",
       "0        207623.571429                 373239.0              85240.928571   \n",
       "1        108131.945625                 245781.0              49156.200000   \n",
       "2        254350.125000                      0.0                  0.000000   \n",
       "3         94518.900000                      0.0                  0.000000   \n",
       "4        219042.000000                 543087.0             189469.500000   \n",
       "\n",
       "   AMT_CREDIT_SUM_LIMIT_MAX  AMT_CREDIT_SUM_LIMIT_MEAN  \\\n",
       "0                     0.000                    0.00000   \n",
       "1                 31988.565                 7997.14125   \n",
       "2                810000.000               202500.00000   \n",
       "3                     0.000                    0.00000   \n",
       "4                     0.000                    0.00000   \n",
       "\n",
       "   AMT_CREDIT_SUM_OVERDUE_MAX  AMT_CREDIT_SUM_OVERDUE_MEAN  AMT_ANNUITY_MAX  \\\n",
       "0                         0.0                          0.0          10822.5   \n",
       "1                         0.0                          0.0              0.0   \n",
       "2                         0.0                          0.0              NaN   \n",
       "3                         0.0                          0.0              NaN   \n",
       "4                         0.0                          0.0           4261.5   \n",
       "\n",
       "   AMT_ANNUITY_MEAN  STATUS_MAX_MAX  STATUS_MEAN_MEAN  \n",
       "0       3545.357143             0.2          0.002198  \n",
       "1          0.000000             0.2          0.059844  \n",
       "2               NaN             NaN               NaN  \n",
       "3               NaN             NaN               NaN  \n",
       "4       1420.500000             0.0          0.000000  "
      ]
     },
     "execution_count": 197,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_bureau = merged_bureau.groupby('SK_ID_CURR').agg({'CREDIT_DAY_OVERDUE':['max', 'mean'],\n",
    "                                                      'AMT_CREDIT_MAX_OVERDUE':['max'],\n",
    "                                                      'CNT_CREDIT_PROLONG':['max'],\n",
    "                                                      'AMT_CREDIT_SUM':['max','mean'],\n",
    "                                                      'AMT_CREDIT_SUM_DEBT':['max','mean'],\n",
    "                                                      'AMT_CREDIT_SUM_LIMIT':['max','mean'],\n",
    "                                                      'AMT_CREDIT_SUM_OVERDUE':['max','mean'],\n",
    "                                                      'AMT_ANNUITY':['max','mean'],\n",
    "                                                      'STATUS_MAX':['max'],'STATUS_MEAN':['mean']})\n",
    "agg_bureau.columns = agg_bureau.columns.map('_'.join).str.strip('_')\n",
    "agg_bureau.columns = map(str.upper, agg_bureau.columns)\n",
    "agg_bureau.reset_index(inplace=True)\n",
    "agg_bureau.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 198,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "agg_bureau['CREDIT_DAY_OVERDUE_MAX'] = ((agg_bureau['CREDIT_DAY_OVERDUE_MAX'] - agg_bureau['CREDIT_DAY_OVERDUE_MAX'].min())/(agg_bureau['CREDIT_DAY_OVERDUE_MAX'].max() - agg_bureau['CREDIT_DAY_OVERDUE_MAX'].min()))\n",
    "agg_bureau['CREDIT_DAY_OVERDUE_MEAN'] = ((agg_bureau['CREDIT_DAY_OVERDUE_MEAN'] - agg_bureau['CREDIT_DAY_OVERDUE_MEAN'].min())/(agg_bureau['CREDIT_DAY_OVERDUE_MEAN'].max() - agg_bureau['CREDIT_DAY_OVERDUE_MEAN'].min()))\n",
    "agg_bureau['AMT_CREDIT_MAX_OVERDUE_MAX'] = ((agg_bureau['AMT_CREDIT_MAX_OVERDUE_MAX'] - agg_bureau['AMT_CREDIT_MAX_OVERDUE_MAX'].min())/(agg_bureau['AMT_CREDIT_MAX_OVERDUE_MAX'].max() - agg_bureau['AMT_CREDIT_MAX_OVERDUE_MAX'].min()))\n",
    "agg_bureau['CNT_CREDIT_PROLONG_MAX'] = ((agg_bureau['CNT_CREDIT_PROLONG_MAX'] - agg_bureau['CNT_CREDIT_PROLONG_MAX'].min())/(agg_bureau['CNT_CREDIT_PROLONG_MAX'].max() - agg_bureau['CNT_CREDIT_PROLONG_MAX'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_MAX'] = ((agg_bureau['AMT_CREDIT_SUM_MAX'] - agg_bureau['AMT_CREDIT_SUM_MAX'].min())/(agg_bureau['AMT_CREDIT_SUM_MAX'].max() - agg_bureau['AMT_CREDIT_SUM_MAX'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_MEAN'] = ((agg_bureau['AMT_CREDIT_SUM_MEAN'] - agg_bureau['AMT_CREDIT_SUM_MEAN'].min())/(agg_bureau['AMT_CREDIT_SUM_MEAN'].max() - agg_bureau['AMT_CREDIT_SUM_MEAN'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_DEBT_MAX'] = ((agg_bureau['AMT_CREDIT_SUM_DEBT_MAX'] - agg_bureau['AMT_CREDIT_SUM_DEBT_MAX'].min())/(agg_bureau['AMT_CREDIT_SUM_DEBT_MAX'].max() - agg_bureau['AMT_CREDIT_SUM_DEBT_MAX'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_DEBT_MEAN'] = ((agg_bureau['AMT_CREDIT_SUM_DEBT_MEAN'] - agg_bureau['AMT_CREDIT_SUM_DEBT_MEAN'].min())/(agg_bureau['AMT_CREDIT_SUM_DEBT_MEAN'].max() - agg_bureau['AMT_CREDIT_SUM_DEBT_MEAN'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_LIMIT_MAX'] = ((agg_bureau['AMT_CREDIT_SUM_LIMIT_MAX'] - agg_bureau['AMT_CREDIT_SUM_LIMIT_MAX'].min())/(agg_bureau['AMT_CREDIT_SUM_LIMIT_MAX'].max() - agg_bureau['AMT_CREDIT_SUM_LIMIT_MAX'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_LIMIT_MEAN'] = ((agg_bureau['AMT_CREDIT_SUM_LIMIT_MEAN'] - agg_bureau['AMT_CREDIT_SUM_LIMIT_MEAN'].min())/(agg_bureau['AMT_CREDIT_SUM_LIMIT_MEAN'].max() - agg_bureau['AMT_CREDIT_SUM_LIMIT_MEAN'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_OVERDUE_MAX'] = ((agg_bureau['AMT_CREDIT_SUM_OVERDUE_MAX'] - agg_bureau['AMT_CREDIT_SUM_OVERDUE_MAX'].min())/(agg_bureau['AMT_CREDIT_SUM_OVERDUE_MAX'].max() - agg_bureau['AMT_CREDIT_SUM_OVERDUE_MAX'].min()))\n",
    "agg_bureau['AMT_CREDIT_SUM_OVERDUE_MEAN'] = ((agg_bureau['AMT_CREDIT_SUM_OVERDUE_MEAN'] - agg_bureau['AMT_CREDIT_SUM_OVERDUE_MEAN'].min())/(agg_bureau['AMT_CREDIT_SUM_OVERDUE_MEAN'].max() - agg_bureau['AMT_CREDIT_SUM_OVERDUE_MEAN'].min()))\n",
    "agg_bureau['AMT_ANNUITY_MAX'] = ((agg_bureau['AMT_ANNUITY_MAX'] - agg_bureau['AMT_ANNUITY_MAX'].min())/(agg_bureau['AMT_ANNUITY_MAX'].max() - agg_bureau['AMT_ANNUITY_MAX'].min()))\n",
    "agg_bureau['AMT_ANNUITY_MEAN'] = ((agg_bureau['AMT_ANNUITY_MEAN'] - agg_bureau['AMT_ANNUITY_MEAN'].min())/(agg_bureau['AMT_ANNUITY_MEAN'].max() - agg_bureau['AMT_ANNUITY_MEAN'].min()))\n",
    "agg_bureau['STATUS_MAX_MAX'] = ((agg_bureau['STATUS_MAX_MAX'] - agg_bureau['STATUS_MAX_MAX'].min())/(agg_bureau['STATUS_MAX_MAX'].max() - agg_bureau['STATUS_MAX_MAX'].min()))\n",
    "agg_bureau['STATUS_MEAN_MEAN'] = ((agg_bureau['STATUS_MEAN_MEAN'] - agg_bureau['STATUS_MEAN_MEAN'].min())/(agg_bureau['STATUS_MEAN_MEAN'].max() - agg_bureau['STATUS_MEAN_MEAN'].min()))\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 199,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_processed_app_train = pd.merge(processed_app_train, agg_bureau, how='left', on='SK_ID_CURR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 200,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>TARGET</th>\n",
       "      <th>FLAG_OWN_CAR</th>\n",
       "      <th>FLAG_OWN_REALTY</th>\n",
       "      <th>CNT_CHILDREN</th>\n",
       "      <th>AMT_INCOME_TOTAL</th>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <th>AMT_GOODS_PRICE</th>\n",
       "      <th>REGION_POPULATION_RELATIVE</th>\n",
       "      <th>...</th>\n",
       "      <th>AMT_CREDIT_SUM_DEBT_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_DEBT_MEAN</th>\n",
       "      <th>AMT_CREDIT_SUM_LIMIT_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_LIMIT_MEAN</th>\n",
       "      <th>AMT_CREDIT_SUM_OVERDUE_MAX</th>\n",
       "      <th>AMT_CREDIT_SUM_OVERDUE_MEAN</th>\n",
       "      <th>AMT_ANNUITY_MAX</th>\n",
       "      <th>AMT_ANNUITY_MEAN</th>\n",
       "      <th>STATUS_MAX_MAX</th>\n",
       "      <th>STATUS_MEAN_MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100002</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.245235</td>\n",
       "      <td>0.489168</td>\n",
       "      <td>0.537543</td>\n",
       "      <td>0.468926</td>\n",
       "      <td>0.256321</td>\n",
       "      <td>...</td>\n",
       "      <td>0.001597</td>\n",
       "      <td>0.02144</td>\n",
       "      <td>0.007167</td>\n",
       "      <td>0.023030</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.2</td>\n",
       "      <td>0.059844</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100003</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.279379</td>\n",
       "      <td>0.746353</td>\n",
       "      <td>0.610134</td>\n",
       "      <td>0.722716</td>\n",
       "      <td>0.045016</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000153</td>\n",
       "      <td>0.02051</td>\n",
       "      <td>0.172443</td>\n",
       "      <td>0.065332</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100004</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.114842</td>\n",
       "      <td>0.244146</td>\n",
       "      <td>0.281841</td>\n",
       "      <td>0.261439</td>\n",
       "      <td>0.134897</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000153</td>\n",
       "      <td>0.02051</td>\n",
       "      <td>0.000371</td>\n",
       "      <td>0.021291</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100006</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.197110</td>\n",
       "      <td>0.430802</td>\n",
       "      <td>0.573785</td>\n",
       "      <td>0.432651</td>\n",
       "      <td>0.107023</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100007</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.184605</td>\n",
       "      <td>0.540826</td>\n",
       "      <td>0.513513</td>\n",
       "      <td>0.551331</td>\n",
       "      <td>0.392880</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000153</td>\n",
       "      <td>0.02051</td>\n",
       "      <td>0.000371</td>\n",
       "      <td>0.021291</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 269 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  TARGET  FLAG_OWN_CAR  FLAG_OWN_REALTY  CNT_CHILDREN  \\\n",
       "0      100002       1             0                1           0.0   \n",
       "1      100003       0             0                0           0.0   \n",
       "2      100004       0             1                1           0.0   \n",
       "3      100006       0             0                1           0.0   \n",
       "4      100007       0             0                1           0.0   \n",
       "\n",
       "   AMT_INCOME_TOTAL  AMT_CREDIT  AMT_ANNUITY  AMT_GOODS_PRICE  \\\n",
       "0          0.245235    0.489168     0.537543         0.468926   \n",
       "1          0.279379    0.746353     0.610134         0.722716   \n",
       "2          0.114842    0.244146     0.281841         0.261439   \n",
       "3          0.197110    0.430802     0.573785         0.432651   \n",
       "4          0.184605    0.540826     0.513513         0.551331   \n",
       "\n",
       "   REGION_POPULATION_RELATIVE        ...         AMT_CREDIT_SUM_DEBT_MAX  \\\n",
       "0                    0.256321        ...                        0.001597   \n",
       "1                    0.045016        ...                        0.000153   \n",
       "2                    0.134897        ...                        0.000153   \n",
       "3                    0.107023        ...                             NaN   \n",
       "4                    0.392880        ...                        0.000153   \n",
       "\n",
       "   AMT_CREDIT_SUM_DEBT_MEAN  AMT_CREDIT_SUM_LIMIT_MAX  \\\n",
       "0                   0.02144                  0.007167   \n",
       "1                   0.02051                  0.172443   \n",
       "2                   0.02051                  0.000371   \n",
       "3                       NaN                       NaN   \n",
       "4                   0.02051                  0.000371   \n",
       "\n",
       "   AMT_CREDIT_SUM_LIMIT_MEAN  AMT_CREDIT_SUM_OVERDUE_MAX  \\\n",
       "0                   0.023030                         0.0   \n",
       "1                   0.065332                         0.0   \n",
       "2                   0.021291                         0.0   \n",
       "3                        NaN                         NaN   \n",
       "4                   0.021291                         0.0   \n",
       "\n",
       "   AMT_CREDIT_SUM_OVERDUE_MEAN  AMT_ANNUITY_MAX  AMT_ANNUITY_MEAN  \\\n",
       "0                          0.0              0.0               0.0   \n",
       "1                          0.0              NaN               NaN   \n",
       "2                          0.0              NaN               NaN   \n",
       "3                          NaN              NaN               NaN   \n",
       "4                          0.0              NaN               NaN   \n",
       "\n",
       "   STATUS_MAX_MAX  STATUS_MEAN_MEAN  \n",
       "0             0.2          0.059844  \n",
       "1             NaN               NaN  \n",
       "2             NaN               NaN  \n",
       "3             NaN               NaN  \n",
       "4             NaN               NaN  \n",
       "\n",
       "[5 rows x 269 columns]"
      ]
     },
     "execution_count": 200,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_processed_app_train.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TABLE POS_CASH_balance.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(10001358, 8)"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pos_cash = pd.read_csv(\"data/POS_CASH_balance.csv\")\n",
    "pos_cash.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 202,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <th>CNT_INSTALMENT</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE</th>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <th>SK_DPD</th>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1803195</td>\n",
       "      <td>182943</td>\n",
       "      <td>-31</td>\n",
       "      <td>48.0</td>\n",
       "      <td>45.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1715348</td>\n",
       "      <td>367990</td>\n",
       "      <td>-33</td>\n",
       "      <td>36.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1784872</td>\n",
       "      <td>397406</td>\n",
       "      <td>-32</td>\n",
       "      <td>12.0</td>\n",
       "      <td>9.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1903291</td>\n",
       "      <td>269225</td>\n",
       "      <td>-35</td>\n",
       "      <td>48.0</td>\n",
       "      <td>42.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2341044</td>\n",
       "      <td>334279</td>\n",
       "      <td>-35</td>\n",
       "      <td>36.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_PREV  SK_ID_CURR  MONTHS_BALANCE  CNT_INSTALMENT  \\\n",
       "0     1803195      182943             -31            48.0   \n",
       "1     1715348      367990             -33            36.0   \n",
       "2     1784872      397406             -32            12.0   \n",
       "3     1903291      269225             -35            48.0   \n",
       "4     2341044      334279             -35            36.0   \n",
       "\n",
       "   CNT_INSTALMENT_FUTURE NAME_CONTRACT_STATUS  SK_DPD  SK_DPD_DEF  \n",
       "0                   45.0               Active       0           0  \n",
       "1                   35.0               Active       0           0  \n",
       "2                    9.0               Active       0           0  \n",
       "3                   42.0               Active       0           0  \n",
       "4                   35.0               Active       0           0  "
      ]
     },
     "execution_count": 202,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pos_cash.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 203,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>CNT_INSTALMENT_MAX</th>\n",
       "      <th>CNT_INSTALMENT_MEAN</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE_MAX</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE_MEAN</th>\n",
       "      <th>SK_DPD_MAX</th>\n",
       "      <th>SK_DPD_MEAN</th>\n",
       "      <th>SK_DPD_DEF_MAX</th>\n",
       "      <th>SK_DPD_DEF_MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>4.0</td>\n",
       "      <td>4.000000</td>\n",
       "      <td>4.0</td>\n",
       "      <td>1.444444</td>\n",
       "      <td>7</td>\n",
       "      <td>0.777778</td>\n",
       "      <td>7</td>\n",
       "      <td>0.777778</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>24.0</td>\n",
       "      <td>24.000000</td>\n",
       "      <td>24.0</td>\n",
       "      <td>15.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>12.0</td>\n",
       "      <td>10.107143</td>\n",
       "      <td>12.0</td>\n",
       "      <td>5.785714</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>4.0</td>\n",
       "      <td>3.750000</td>\n",
       "      <td>4.0</td>\n",
       "      <td>2.250000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>12.0</td>\n",
       "      <td>11.700000</td>\n",
       "      <td>12.0</td>\n",
       "      <td>7.200000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  CNT_INSTALMENT_MAX  CNT_INSTALMENT_MEAN  \\\n",
       "0      100001                 4.0             4.000000   \n",
       "1      100002                24.0            24.000000   \n",
       "2      100003                12.0            10.107143   \n",
       "3      100004                 4.0             3.750000   \n",
       "4      100005                12.0            11.700000   \n",
       "\n",
       "   CNT_INSTALMENT_FUTURE_MAX  CNT_INSTALMENT_FUTURE_MEAN  SK_DPD_MAX  \\\n",
       "0                        4.0                    1.444444           7   \n",
       "1                       24.0                   15.000000           0   \n",
       "2                       12.0                    5.785714           0   \n",
       "3                        4.0                    2.250000           0   \n",
       "4                       12.0                    7.200000           0   \n",
       "\n",
       "   SK_DPD_MEAN  SK_DPD_DEF_MAX  SK_DPD_DEF_MEAN  \n",
       "0     0.777778               7         0.777778  \n",
       "1     0.000000               0         0.000000  \n",
       "2     0.000000               0         0.000000  \n",
       "3     0.000000               0         0.000000  \n",
       "4     0.000000               0         0.000000  "
      ]
     },
     "execution_count": 203,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_pos_cash = pos_cash.groupby('SK_ID_CURR').agg({'CNT_INSTALMENT':['max', 'mean'],\n",
    "                                                      'CNT_INSTALMENT_FUTURE':['max', 'mean'],\n",
    "                                                      'SK_DPD':['max', 'mean'],\n",
    "                                                      'SK_DPD_DEF':['max', 'mean']})\n",
    "agg_pos_cash.columns = agg_pos_cash.columns.map('_'.join).str.strip('_')\n",
    "agg_pos_cash.columns = map(str.upper, agg_pos_cash.columns)\n",
    "agg_pos_cash.reset_index(inplace=True)\n",
    "agg_pos_cash.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 204,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "agg_pos_cash['CNT_INSTALMENT_MAX'] = ((agg_pos_cash['CNT_INSTALMENT_MAX'] - agg_pos_cash['CNT_INSTALMENT_MAX'].min())/(agg_pos_cash['CNT_INSTALMENT_MAX'].max() - agg_pos_cash['CNT_INSTALMENT_MAX'].min()))\n",
    "agg_pos_cash['CNT_INSTALMENT_MEAN'] = ((agg_pos_cash['CNT_INSTALMENT_MEAN'] - agg_pos_cash['CNT_INSTALMENT_MEAN'].min())/(agg_pos_cash['CNT_INSTALMENT_MEAN'].max() - agg_pos_cash['CNT_INSTALMENT_MEAN'].min()))\n",
    "agg_pos_cash['CNT_INSTALMENT_FUTURE_MAX'] = ((agg_pos_cash['CNT_INSTALMENT_FUTURE_MAX'] - agg_pos_cash['CNT_INSTALMENT_FUTURE_MAX'].min())/(agg_pos_cash['CNT_INSTALMENT_FUTURE_MAX'].max() - agg_pos_cash['CNT_INSTALMENT_FUTURE_MAX'].min()))\n",
    "agg_pos_cash['CNT_INSTALMENT_FUTURE_MEAN'] = ((agg_pos_cash['CNT_INSTALMENT_FUTURE_MEAN'] - agg_pos_cash['CNT_INSTALMENT_FUTURE_MEAN'].min())/(agg_pos_cash['CNT_INSTALMENT_FUTURE_MEAN'].max() - agg_pos_cash['CNT_INSTALMENT_FUTURE_MEAN'].min()))\n",
    "agg_pos_cash['SK_DPD_MAX'] = ((agg_pos_cash['SK_DPD_MAX'] - agg_pos_cash['SK_DPD_MAX'].min())/(agg_pos_cash['SK_DPD_MAX'].max() - agg_pos_cash['SK_DPD_MAX'].min()))\n",
    "agg_pos_cash['SK_DPD_MEAN'] = ((agg_pos_cash['SK_DPD_MEAN'] - agg_pos_cash['SK_DPD_MEAN'].min())/(agg_pos_cash['SK_DPD_MEAN'].max() - agg_pos_cash['SK_DPD_MEAN'].min()))\n",
    "agg_pos_cash['SK_DPD_DEF_MAX'] = ((agg_pos_cash['SK_DPD_DEF_MAX'] - agg_pos_cash['SK_DPD_DEF_MAX'].min())/(agg_pos_cash['SK_DPD_DEF_MAX'].max() - agg_pos_cash['SK_DPD_DEF_MAX'].min()))\n",
    "agg_pos_cash['SK_DPD_DEF_MEAN'] = ((agg_pos_cash['SK_DPD_DEF_MEAN'] - agg_pos_cash['SK_DPD_DEF_MEAN'].min())/(agg_pos_cash['SK_DPD_DEF_MEAN'].max() - agg_pos_cash['SK_DPD_DEF_MEAN'].min()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 205,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_processed_app_train = pd.merge(merged_processed_app_train, agg_pos_cash, how='left', on='SK_ID_CURR')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TABLE credit_card_balance.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(3840312, 23)"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "credit_card = pd.read_csv(\"data/credit_card_balance.csv\")\n",
    "credit_card.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 207,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <th>AMT_BALANCE</th>\n",
       "      <th>AMT_CREDIT_LIMIT_ACTUAL</th>\n",
       "      <th>AMT_DRAWINGS_ATM_CURRENT</th>\n",
       "      <th>AMT_DRAWINGS_CURRENT</th>\n",
       "      <th>AMT_DRAWINGS_OTHER_CURRENT</th>\n",
       "      <th>AMT_DRAWINGS_POS_CURRENT</th>\n",
       "      <th>AMT_INST_MIN_REGULARITY</th>\n",
       "      <th>...</th>\n",
       "      <th>AMT_RECIVABLE</th>\n",
       "      <th>AMT_TOTAL_RECEIVABLE</th>\n",
       "      <th>CNT_DRAWINGS_ATM_CURRENT</th>\n",
       "      <th>CNT_DRAWINGS_CURRENT</th>\n",
       "      <th>CNT_DRAWINGS_OTHER_CURRENT</th>\n",
       "      <th>CNT_DRAWINGS_POS_CURRENT</th>\n",
       "      <th>CNT_INSTALMENT_MATURE_CUM</th>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <th>SK_DPD</th>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2562384</td>\n",
       "      <td>378907</td>\n",
       "      <td>-6</td>\n",
       "      <td>56.970</td>\n",
       "      <td>135000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>877.5</td>\n",
       "      <td>0.0</td>\n",
       "      <td>877.5</td>\n",
       "      <td>1700.325</td>\n",
       "      <td>...</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>35.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2582071</td>\n",
       "      <td>363914</td>\n",
       "      <td>-1</td>\n",
       "      <td>63975.555</td>\n",
       "      <td>45000</td>\n",
       "      <td>2250.0</td>\n",
       "      <td>2250.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2250.000</td>\n",
       "      <td>...</td>\n",
       "      <td>64875.555</td>\n",
       "      <td>64875.555</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>69.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1740877</td>\n",
       "      <td>371185</td>\n",
       "      <td>-7</td>\n",
       "      <td>31815.225</td>\n",
       "      <td>450000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>2250.000</td>\n",
       "      <td>...</td>\n",
       "      <td>31460.085</td>\n",
       "      <td>31460.085</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>30.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1389973</td>\n",
       "      <td>337855</td>\n",
       "      <td>-4</td>\n",
       "      <td>236572.110</td>\n",
       "      <td>225000</td>\n",
       "      <td>2250.0</td>\n",
       "      <td>2250.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11795.760</td>\n",
       "      <td>...</td>\n",
       "      <td>233048.970</td>\n",
       "      <td>233048.970</td>\n",
       "      <td>1.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>10.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1891521</td>\n",
       "      <td>126868</td>\n",
       "      <td>-1</td>\n",
       "      <td>453919.455</td>\n",
       "      <td>450000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11547.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>11547.0</td>\n",
       "      <td>22924.890</td>\n",
       "      <td>...</td>\n",
       "      <td>453919.455</td>\n",
       "      <td>453919.455</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1</td>\n",
       "      <td>0.0</td>\n",
       "      <td>1.0</td>\n",
       "      <td>101.0</td>\n",
       "      <td>Active</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_PREV  SK_ID_CURR  MONTHS_BALANCE  AMT_BALANCE  \\\n",
       "0     2562384      378907              -6       56.970   \n",
       "1     2582071      363914              -1    63975.555   \n",
       "2     1740877      371185              -7    31815.225   \n",
       "3     1389973      337855              -4   236572.110   \n",
       "4     1891521      126868              -1   453919.455   \n",
       "\n",
       "   AMT_CREDIT_LIMIT_ACTUAL  AMT_DRAWINGS_ATM_CURRENT  AMT_DRAWINGS_CURRENT  \\\n",
       "0                   135000                       0.0                 877.5   \n",
       "1                    45000                    2250.0                2250.0   \n",
       "2                   450000                       0.0                   0.0   \n",
       "3                   225000                    2250.0                2250.0   \n",
       "4                   450000                       0.0               11547.0   \n",
       "\n",
       "   AMT_DRAWINGS_OTHER_CURRENT  AMT_DRAWINGS_POS_CURRENT  \\\n",
       "0                         0.0                     877.5   \n",
       "1                         0.0                       0.0   \n",
       "2                         0.0                       0.0   \n",
       "3                         0.0                       0.0   \n",
       "4                         0.0                   11547.0   \n",
       "\n",
       "   AMT_INST_MIN_REGULARITY     ...      AMT_RECIVABLE  AMT_TOTAL_RECEIVABLE  \\\n",
       "0                 1700.325     ...              0.000                 0.000   \n",
       "1                 2250.000     ...          64875.555             64875.555   \n",
       "2                 2250.000     ...          31460.085             31460.085   \n",
       "3                11795.760     ...         233048.970            233048.970   \n",
       "4                22924.890     ...         453919.455            453919.455   \n",
       "\n",
       "   CNT_DRAWINGS_ATM_CURRENT  CNT_DRAWINGS_CURRENT  CNT_DRAWINGS_OTHER_CURRENT  \\\n",
       "0                       0.0                     1                         0.0   \n",
       "1                       1.0                     1                         0.0   \n",
       "2                       0.0                     0                         0.0   \n",
       "3                       1.0                     1                         0.0   \n",
       "4                       0.0                     1                         0.0   \n",
       "\n",
       "   CNT_DRAWINGS_POS_CURRENT  CNT_INSTALMENT_MATURE_CUM  NAME_CONTRACT_STATUS  \\\n",
       "0                       1.0                       35.0                Active   \n",
       "1                       0.0                       69.0                Active   \n",
       "2                       0.0                       30.0                Active   \n",
       "3                       0.0                       10.0                Active   \n",
       "4                       1.0                      101.0                Active   \n",
       "\n",
       "   SK_DPD  SK_DPD_DEF  \n",
       "0       0           0  \n",
       "1       0           0  \n",
       "2       0           0  \n",
       "3       0           0  \n",
       "4       0           0  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 207,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "credit_card.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 208,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>AMT_BALANCE_MAX</th>\n",
       "      <th>AMT_BALANCE_MEAN</th>\n",
       "      <th>AMT_CREDIT_LIMIT_ACTUAL_MAX</th>\n",
       "      <th>AMT_CREDIT_LIMIT_ACTUAL_MEAN</th>\n",
       "      <th>CNT_DRAWINGS_ATM_CURRENT_MAX</th>\n",
       "      <th>CNT_DRAWINGS_ATM_CURRENT_MEAN</th>\n",
       "      <th>CNT_INSTALMENT_MATURE_CUM_MAX</th>\n",
       "      <th>CNT_INSTALMENT_MATURE_CUM_MEAN</th>\n",
       "      <th>SK_DPD_MAX</th>\n",
       "      <th>SK_DPD_MEAN</th>\n",
       "      <th>SK_DPD_DEF_MAX</th>\n",
       "      <th>SK_DPD_DEF_MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100006</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>270000</td>\n",
       "      <td>270000.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100011</td>\n",
       "      <td>189000.00</td>\n",
       "      <td>54482.111149</td>\n",
       "      <td>180000</td>\n",
       "      <td>164189.189189</td>\n",
       "      <td>4.0</td>\n",
       "      <td>0.054054</td>\n",
       "      <td>33.0</td>\n",
       "      <td>25.767123</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100013</td>\n",
       "      <td>161420.22</td>\n",
       "      <td>18159.919219</td>\n",
       "      <td>157500</td>\n",
       "      <td>131718.750000</td>\n",
       "      <td>7.0</td>\n",
       "      <td>0.255556</td>\n",
       "      <td>22.0</td>\n",
       "      <td>18.719101</td>\n",
       "      <td>1</td>\n",
       "      <td>0.010417</td>\n",
       "      <td>1</td>\n",
       "      <td>0.010417</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100021</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>675000</td>\n",
       "      <td>675000.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100023</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>225000</td>\n",
       "      <td>135000.000000</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "      <td>0</td>\n",
       "      <td>0.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  AMT_BALANCE_MAX  AMT_BALANCE_MEAN  AMT_CREDIT_LIMIT_ACTUAL_MAX  \\\n",
       "0      100006             0.00          0.000000                       270000   \n",
       "1      100011        189000.00      54482.111149                       180000   \n",
       "2      100013        161420.22      18159.919219                       157500   \n",
       "3      100021             0.00          0.000000                       675000   \n",
       "4      100023             0.00          0.000000                       225000   \n",
       "\n",
       "   AMT_CREDIT_LIMIT_ACTUAL_MEAN  CNT_DRAWINGS_ATM_CURRENT_MAX  \\\n",
       "0                 270000.000000                           NaN   \n",
       "1                 164189.189189                           4.0   \n",
       "2                 131718.750000                           7.0   \n",
       "3                 675000.000000                           NaN   \n",
       "4                 135000.000000                           NaN   \n",
       "\n",
       "   CNT_DRAWINGS_ATM_CURRENT_MEAN  CNT_INSTALMENT_MATURE_CUM_MAX  \\\n",
       "0                            NaN                            0.0   \n",
       "1                       0.054054                           33.0   \n",
       "2                       0.255556                           22.0   \n",
       "3                            NaN                            0.0   \n",
       "4                            NaN                            0.0   \n",
       "\n",
       "   CNT_INSTALMENT_MATURE_CUM_MEAN  SK_DPD_MAX  SK_DPD_MEAN  SK_DPD_DEF_MAX  \\\n",
       "0                        0.000000           0     0.000000               0   \n",
       "1                       25.767123           0     0.000000               0   \n",
       "2                       18.719101           1     0.010417               1   \n",
       "3                        0.000000           0     0.000000               0   \n",
       "4                        0.000000           0     0.000000               0   \n",
       "\n",
       "   SK_DPD_DEF_MEAN  \n",
       "0         0.000000  \n",
       "1         0.000000  \n",
       "2         0.010417  \n",
       "3         0.000000  \n",
       "4         0.000000  "
      ]
     },
     "execution_count": 208,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_credit_card = credit_card.groupby('SK_ID_CURR').agg({'AMT_BALANCE':['max', 'mean'],\n",
    "                                                     'AMT_CREDIT_LIMIT_ACTUAL':['max', 'mean'],\n",
    "                                                     'CNT_DRAWINGS_ATM_CURRENT':['max', 'mean'],\n",
    "                                                     'CNT_INSTALMENT_MATURE_CUM':['max', 'mean'],\n",
    "                                                     'SK_DPD':['max', 'mean'],\n",
    "                                                     'SK_DPD_DEF':['max', 'mean']})\n",
    "agg_credit_card.columns = agg_credit_card.columns.map('_'.join).str.strip('_')\n",
    "agg_credit_card.columns = map(str.upper, agg_credit_card.columns)\n",
    "agg_credit_card.reset_index(inplace=True)\n",
    "agg_credit_card.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 209,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "agg_credit_card['AMT_BALANCE_MAX'] = ((agg_credit_card['AMT_BALANCE_MAX'] - agg_credit_card['AMT_BALANCE_MAX'].min())/(agg_credit_card['AMT_BALANCE_MAX'].max() - agg_credit_card['AMT_BALANCE_MAX'].min()))\n",
    "agg_credit_card['AMT_BALANCE_MEAN'] = ((agg_credit_card['AMT_BALANCE_MEAN'] - agg_credit_card['AMT_BALANCE_MEAN'].min())/(agg_credit_card['AMT_BALANCE_MEAN'].max() - agg_credit_card['AMT_BALANCE_MEAN'].min()))\n",
    "agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MAX'] = ((agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MAX'] - agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MAX'].min())/(agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MAX'].max() - agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MAX'].min()))\n",
    "agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MEAN'] = ((agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MEAN'] - agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MEAN'].min())/(agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MEAN'].max() - agg_credit_card['AMT_CREDIT_LIMIT_ACTUAL_MEAN'].min()))\n",
    "agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MAX'] = ((agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MAX'] - agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MAX'].min())/(agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MAX'].max() - agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MAX'].min()))\n",
    "agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MEAN'] = ((agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MEAN'] - agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MEAN'].min())/(agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MEAN'].max() - agg_credit_card['CNT_DRAWINGS_ATM_CURRENT_MEAN'].min()))\n",
    "agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MAX'] = ((agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MAX'] - agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MAX'].min())/(agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MAX'].max() - agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MAX'].min()))\n",
    "agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MEAN'] = ((agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MEAN'] - agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MEAN'].min())/(agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MEAN'].max() - agg_credit_card['CNT_INSTALMENT_MATURE_CUM_MEAN'].min()))\n",
    "agg_credit_card['SK_DPD_MAX'] = ((agg_credit_card['SK_DPD_MAX'] - agg_credit_card['SK_DPD_MAX'].min())/(agg_credit_card['SK_DPD_MAX'].max() - agg_credit_card['SK_DPD_MAX'].min()))\n",
    "agg_credit_card['SK_DPD_MEAN'] = ((agg_credit_card['SK_DPD_MEAN'] - agg_credit_card['SK_DPD_MEAN'].min())/(agg_credit_card['SK_DPD_MEAN'].max() - agg_credit_card['SK_DPD_MEAN'].min()))\n",
    "agg_credit_card['SK_DPD_DEF_MAX'] = ((agg_credit_card['SK_DPD_DEF_MAX'] - agg_credit_card['SK_DPD_DEF_MAX'].min())/(agg_credit_card['SK_DPD_DEF_MAX'].max() - agg_credit_card['SK_DPD_DEF_MAX'].min()))\n",
    "agg_credit_card['SK_DPD_DEF_MEAN'] = ((agg_credit_card['SK_DPD_DEF_MEAN'] - agg_credit_card['SK_DPD_DEF_MEAN'].min())/(agg_credit_card['SK_DPD_DEF_MEAN'].max() - agg_credit_card['SK_DPD_DEF_MEAN'].min()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 210,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_processed_app_train = pd.merge(merged_processed_app_train, agg_credit_card, how='left', on='SK_ID_CURR')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TABLE previous_application.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1670214, 37)"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application = pd.read_csv(\"data/previous_application.csv\")\n",
    "previous_application.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 212,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>NAME_CONTRACT_TYPE</th>\n",
       "      <th>AMT_ANNUITY</th>\n",
       "      <th>AMT_APPLICATION</th>\n",
       "      <th>AMT_CREDIT</th>\n",
       "      <th>AMT_DOWN_PAYMENT</th>\n",
       "      <th>AMT_GOODS_PRICE</th>\n",
       "      <th>WEEKDAY_APPR_PROCESS_START</th>\n",
       "      <th>HOUR_APPR_PROCESS_START</th>\n",
       "      <th>...</th>\n",
       "      <th>NAME_SELLER_INDUSTRY</th>\n",
       "      <th>CNT_PAYMENT</th>\n",
       "      <th>NAME_YIELD_GROUP</th>\n",
       "      <th>PRODUCT_COMBINATION</th>\n",
       "      <th>DAYS_FIRST_DRAWING</th>\n",
       "      <th>DAYS_FIRST_DUE</th>\n",
       "      <th>DAYS_LAST_DUE_1ST_VERSION</th>\n",
       "      <th>DAYS_LAST_DUE</th>\n",
       "      <th>DAYS_TERMINATION</th>\n",
       "      <th>NFLAG_INSURED_ON_APPROVAL</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2030495</td>\n",
       "      <td>271877</td>\n",
       "      <td>Consumer loans</td>\n",
       "      <td>1730.430</td>\n",
       "      <td>17145.0</td>\n",
       "      <td>17145.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>17145.0</td>\n",
       "      <td>SATURDAY</td>\n",
       "      <td>15</td>\n",
       "      <td>...</td>\n",
       "      <td>Connectivity</td>\n",
       "      <td>12.0</td>\n",
       "      <td>middle</td>\n",
       "      <td>POS mobile with interest</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>-42.0</td>\n",
       "      <td>300.0</td>\n",
       "      <td>-42.0</td>\n",
       "      <td>-37.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2802425</td>\n",
       "      <td>108129</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>25188.615</td>\n",
       "      <td>607500.0</td>\n",
       "      <td>679671.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>607500.0</td>\n",
       "      <td>THURSDAY</td>\n",
       "      <td>11</td>\n",
       "      <td>...</td>\n",
       "      <td>XNA</td>\n",
       "      <td>36.0</td>\n",
       "      <td>low_action</td>\n",
       "      <td>Cash X-Sell: low</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>-134.0</td>\n",
       "      <td>916.0</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2523466</td>\n",
       "      <td>122040</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>15060.735</td>\n",
       "      <td>112500.0</td>\n",
       "      <td>136444.5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>112500.0</td>\n",
       "      <td>TUESDAY</td>\n",
       "      <td>11</td>\n",
       "      <td>...</td>\n",
       "      <td>XNA</td>\n",
       "      <td>12.0</td>\n",
       "      <td>high</td>\n",
       "      <td>Cash X-Sell: high</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>-271.0</td>\n",
       "      <td>59.0</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2819243</td>\n",
       "      <td>176158</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>47041.335</td>\n",
       "      <td>450000.0</td>\n",
       "      <td>470790.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>450000.0</td>\n",
       "      <td>MONDAY</td>\n",
       "      <td>7</td>\n",
       "      <td>...</td>\n",
       "      <td>XNA</td>\n",
       "      <td>12.0</td>\n",
       "      <td>middle</td>\n",
       "      <td>Cash X-Sell: middle</td>\n",
       "      <td>365243.0</td>\n",
       "      <td>-482.0</td>\n",
       "      <td>-152.0</td>\n",
       "      <td>-182.0</td>\n",
       "      <td>-177.0</td>\n",
       "      <td>1.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>1784265</td>\n",
       "      <td>202054</td>\n",
       "      <td>Cash loans</td>\n",
       "      <td>31924.395</td>\n",
       "      <td>337500.0</td>\n",
       "      <td>404055.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>337500.0</td>\n",
       "      <td>THURSDAY</td>\n",
       "      <td>9</td>\n",
       "      <td>...</td>\n",
       "      <td>XNA</td>\n",
       "      <td>24.0</td>\n",
       "      <td>high</td>\n",
       "      <td>Cash Street: high</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 37 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_PREV  SK_ID_CURR NAME_CONTRACT_TYPE  AMT_ANNUITY  AMT_APPLICATION  \\\n",
       "0     2030495      271877     Consumer loans     1730.430          17145.0   \n",
       "1     2802425      108129         Cash loans    25188.615         607500.0   \n",
       "2     2523466      122040         Cash loans    15060.735         112500.0   \n",
       "3     2819243      176158         Cash loans    47041.335         450000.0   \n",
       "4     1784265      202054         Cash loans    31924.395         337500.0   \n",
       "\n",
       "   AMT_CREDIT  AMT_DOWN_PAYMENT  AMT_GOODS_PRICE WEEKDAY_APPR_PROCESS_START  \\\n",
       "0     17145.0               0.0          17145.0                   SATURDAY   \n",
       "1    679671.0               NaN         607500.0                   THURSDAY   \n",
       "2    136444.5               NaN         112500.0                    TUESDAY   \n",
       "3    470790.0               NaN         450000.0                     MONDAY   \n",
       "4    404055.0               NaN         337500.0                   THURSDAY   \n",
       "\n",
       "   HOUR_APPR_PROCESS_START            ...            NAME_SELLER_INDUSTRY  \\\n",
       "0                       15            ...                    Connectivity   \n",
       "1                       11            ...                             XNA   \n",
       "2                       11            ...                             XNA   \n",
       "3                        7            ...                             XNA   \n",
       "4                        9            ...                             XNA   \n",
       "\n",
       "   CNT_PAYMENT  NAME_YIELD_GROUP       PRODUCT_COMBINATION  \\\n",
       "0         12.0            middle  POS mobile with interest   \n",
       "1         36.0        low_action          Cash X-Sell: low   \n",
       "2         12.0              high         Cash X-Sell: high   \n",
       "3         12.0            middle       Cash X-Sell: middle   \n",
       "4         24.0              high         Cash Street: high   \n",
       "\n",
       "   DAYS_FIRST_DRAWING DAYS_FIRST_DUE DAYS_LAST_DUE_1ST_VERSION  DAYS_LAST_DUE  \\\n",
       "0            365243.0          -42.0                     300.0          -42.0   \n",
       "1            365243.0         -134.0                     916.0       365243.0   \n",
       "2            365243.0         -271.0                      59.0       365243.0   \n",
       "3            365243.0         -482.0                    -152.0         -182.0   \n",
       "4                 NaN            NaN                       NaN            NaN   \n",
       "\n",
       "  DAYS_TERMINATION NFLAG_INSURED_ON_APPROVAL  \n",
       "0            -37.0                       0.0  \n",
       "1         365243.0                       1.0  \n",
       "2         365243.0                       1.0  \n",
       "3           -177.0                       1.0  \n",
       "4              NaN                       NaN  \n",
       "\n",
       "[5 rows x 37 columns]"
      ]
     },
     "execution_count": 212,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 213,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>AMT_ANNUITY_MAX</th>\n",
       "      <th>AMT_ANNUITY_MEAN</th>\n",
       "      <th>AMT_APPLICATION_MAX</th>\n",
       "      <th>AMT_APPLICATION_MEAN</th>\n",
       "      <th>AMT_DOWN_PAYMENT_MAX</th>\n",
       "      <th>AMT_DOWN_PAYMENT_MEAN</th>\n",
       "      <th>RATE_INTEREST_PRIMARY_MAX</th>\n",
       "      <th>RATE_INTEREST_PRIMARY_MEAN</th>\n",
       "      <th>RATE_INTEREST_PRIVILEGED_MAX</th>\n",
       "      <th>RATE_INTEREST_PRIVILEGED_MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>3951.000</td>\n",
       "      <td>3951.000</td>\n",
       "      <td>24835.5</td>\n",
       "      <td>24835.50</td>\n",
       "      <td>2520.0</td>\n",
       "      <td>2520.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>9251.775</td>\n",
       "      <td>9251.775</td>\n",
       "      <td>179055.0</td>\n",
       "      <td>179055.00</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>98356.995</td>\n",
       "      <td>56553.990</td>\n",
       "      <td>900000.0</td>\n",
       "      <td>435436.50</td>\n",
       "      <td>6885.0</td>\n",
       "      <td>3442.5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>5357.250</td>\n",
       "      <td>5357.250</td>\n",
       "      <td>24282.0</td>\n",
       "      <td>24282.00</td>\n",
       "      <td>4860.0</td>\n",
       "      <td>4860.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>4813.200</td>\n",
       "      <td>4813.200</td>\n",
       "      <td>44617.5</td>\n",
       "      <td>22308.75</td>\n",
       "      <td>4464.0</td>\n",
       "      <td>4464.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  AMT_ANNUITY_MAX  AMT_ANNUITY_MEAN  AMT_APPLICATION_MAX  \\\n",
       "0      100001         3951.000          3951.000              24835.5   \n",
       "1      100002         9251.775          9251.775             179055.0   \n",
       "2      100003        98356.995         56553.990             900000.0   \n",
       "3      100004         5357.250          5357.250              24282.0   \n",
       "4      100005         4813.200          4813.200              44617.5   \n",
       "\n",
       "   AMT_APPLICATION_MEAN  AMT_DOWN_PAYMENT_MAX  AMT_DOWN_PAYMENT_MEAN  \\\n",
       "0              24835.50                2520.0                 2520.0   \n",
       "1             179055.00                   0.0                    0.0   \n",
       "2             435436.50                6885.0                 3442.5   \n",
       "3              24282.00                4860.0                 4860.0   \n",
       "4              22308.75                4464.0                 4464.0   \n",
       "\n",
       "   RATE_INTEREST_PRIMARY_MAX  RATE_INTEREST_PRIMARY_MEAN  \\\n",
       "0                        NaN                         NaN   \n",
       "1                        NaN                         NaN   \n",
       "2                        NaN                         NaN   \n",
       "3                        NaN                         NaN   \n",
       "4                        NaN                         NaN   \n",
       "\n",
       "   RATE_INTEREST_PRIVILEGED_MAX  RATE_INTEREST_PRIVILEGED_MEAN  \n",
       "0                           NaN                            NaN  \n",
       "1                           NaN                            NaN  \n",
       "2                           NaN                            NaN  \n",
       "3                           NaN                            NaN  \n",
       "4                           NaN                            NaN  "
      ]
     },
     "execution_count": 213,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_previous_app = previous_application.groupby('SK_ID_CURR').agg({'AMT_ANNUITY':['max', 'mean'],\n",
    "                                                     'AMT_APPLICATION':['max', 'mean'],\n",
    "                                                     'AMT_DOWN_PAYMENT':['max', 'mean'],\n",
    "                                                     'RATE_INTEREST_PRIMARY':['max', 'mean'],\n",
    "                                                     'RATE_INTEREST_PRIVILEGED':['max', 'mean']})\n",
    "agg_previous_app.columns = agg_previous_app.columns.map('_'.join).str.strip('_')\n",
    "agg_previous_app.columns = map(str.upper, agg_previous_app.columns)\n",
    "agg_previous_app.reset_index(inplace=True)\n",
    "agg_previous_app.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 214,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "agg_previous_app['AMT_ANNUITY_MAX'] = ((agg_previous_app['AMT_ANNUITY_MAX'] - agg_previous_app['AMT_ANNUITY_MAX'].min())/(agg_previous_app['AMT_ANNUITY_MAX'].max() - agg_previous_app['AMT_ANNUITY_MAX'].min()))\n",
    "agg_previous_app['AMT_ANNUITY_MEAN'] = ((agg_previous_app['AMT_ANNUITY_MEAN'] - agg_previous_app['AMT_ANNUITY_MEAN'].min())/(agg_previous_app['AMT_ANNUITY_MEAN'].max() - agg_previous_app['AMT_ANNUITY_MEAN'].min()))\n",
    "agg_previous_app['AMT_APPLICATION_MAX'] = ((agg_previous_app['AMT_APPLICATION_MAX'] - agg_previous_app['AMT_APPLICATION_MAX'].min())/(agg_previous_app['AMT_APPLICATION_MAX'].max() - agg_previous_app['AMT_APPLICATION_MAX'].min()))\n",
    "agg_previous_app['AMT_APPLICATION_MEAN'] = ((agg_previous_app['AMT_APPLICATION_MEAN'] - agg_previous_app['AMT_APPLICATION_MEAN'].min())/(agg_previous_app['AMT_APPLICATION_MEAN'].max() - agg_previous_app['AMT_APPLICATION_MEAN'].min()))\n",
    "agg_previous_app['AMT_DOWN_PAYMENT_MAX'] = ((agg_previous_app['AMT_DOWN_PAYMENT_MAX'] - agg_previous_app['AMT_DOWN_PAYMENT_MAX'].min())/(agg_previous_app['AMT_DOWN_PAYMENT_MAX'].max() - agg_previous_app['AMT_DOWN_PAYMENT_MAX'].min()))\n",
    "agg_previous_app['AMT_DOWN_PAYMENT_MEAN'] = ((agg_previous_app['AMT_DOWN_PAYMENT_MEAN'] - agg_previous_app['AMT_DOWN_PAYMENT_MEAN'].min())/(agg_previous_app['AMT_DOWN_PAYMENT_MEAN'].max() - agg_previous_app['AMT_DOWN_PAYMENT_MEAN'].min()))\n",
    "agg_previous_app['RATE_INTEREST_PRIMARY_MAX'] = ((agg_previous_app['RATE_INTEREST_PRIMARY_MAX'] - agg_previous_app['RATE_INTEREST_PRIMARY_MAX'].min())/(agg_previous_app['RATE_INTEREST_PRIMARY_MAX'].max() - agg_previous_app['RATE_INTEREST_PRIMARY_MAX'].min()))\n",
    "agg_previous_app['RATE_INTEREST_PRIMARY_MEAN'] = ((agg_previous_app['RATE_INTEREST_PRIMARY_MEAN'] - agg_previous_app['RATE_INTEREST_PRIMARY_MEAN'].min())/(agg_previous_app['RATE_INTEREST_PRIMARY_MEAN'].max() - agg_previous_app['RATE_INTEREST_PRIMARY_MEAN'].min()))\n",
    "agg_previous_app['RATE_INTEREST_PRIVILEGED_MAX'] = ((agg_previous_app['RATE_INTEREST_PRIVILEGED_MAX'] - agg_previous_app['RATE_INTEREST_PRIVILEGED_MAX'].min())/(agg_previous_app['RATE_INTEREST_PRIVILEGED_MAX'].max() - agg_previous_app['RATE_INTEREST_PRIVILEGED_MAX'].min()))\n",
    "agg_previous_app['RATE_INTEREST_PRIVILEGED_MEAN'] = ((agg_previous_app['RATE_INTEREST_PRIVILEGED_MEAN'] - agg_previous_app['RATE_INTEREST_PRIVILEGED_MEAN'].min())/(agg_previous_app['RATE_INTEREST_PRIVILEGED_MEAN'].max() - agg_previous_app['RATE_INTEREST_PRIVILEGED_MEAN'].min()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 215,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_processed_app_train = pd.merge(merged_processed_app_train, agg_previous_app, how='left', on='SK_ID_CURR')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 216,
   "metadata": {},
   "outputs": [],
   "source": [
    "# get approved apps count\n",
    "approved_previous_app = previous_application[previous_application['NAME_CONTRACT_STATUS'] == 'Approved'].groupby('SK_ID_CURR').size().reset_index(name='APPROVED_COUNT')\n",
    "approved_previous_app['APPROVED_COUNT'] = ((approved_previous_app['APPROVED_COUNT'] - approved_previous_app['APPROVED_COUNT'].min())/(approved_previous_app['APPROVED_COUNT'].max() - approved_previous_app['APPROVED_COUNT'].min()))\n",
    "merged_processed_app_train = pd.merge(merged_processed_app_train, approved_previous_app, how='left', on='SK_ID_CURR')\n",
    "\n",
    "# get refused apps count\n",
    "refused_previous_app = previous_application[previous_application['NAME_CONTRACT_STATUS'] == 'Refused'].groupby('SK_ID_CURR').size().reset_index(name='REFUSED_COUNT')\n",
    "refused_previous_app['REFUSED_COUNT'] = ((refused_previous_app['REFUSED_COUNT'] - refused_previous_app['REFUSED_COUNT'].min())/(refused_previous_app['REFUSED_COUNT'].max() - refused_previous_app['REFUSED_COUNT'].min()))\n",
    "merged_processed_app_train = pd.merge(merged_processed_app_train, refused_previous_app, how='left', on='SK_ID_CURR')\n",
    "\n",
    "# include approved count minus refused count\n",
    "merged_processed_app_train['DIFF_APPROVED_REFUSED'] = merged_processed_app_train['APPROVED_COUNT'] - merged_processed_app_train['REFUSED_COUNT']\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 217,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    5951.000000\n",
       "mean        0.188357\n",
       "std         0.087671\n",
       "min         0.034781\n",
       "25%         0.160716\n",
       "50%         0.189122\n",
       "75%         0.193330\n",
       "max         1.000000\n",
       "Name: RATE_INTEREST_PRIMARY, dtype: float64"
      ]
     },
     "execution_count": 217,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "previous_application['RATE_INTEREST_PRIMARY'].describe()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## TABLE installments_payments.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(13605401, 8)"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "installments_payments = pd.read_csv(\"data/installments_payments.csv\")\n",
    "installments_payments.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 219,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>NUM_INSTALMENT_VERSION</th>\n",
       "      <th>NUM_INSTALMENT_NUMBER</th>\n",
       "      <th>DAYS_INSTALMENT</th>\n",
       "      <th>DAYS_ENTRY_PAYMENT</th>\n",
       "      <th>AMT_INSTALMENT</th>\n",
       "      <th>AMT_PAYMENT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1054186</td>\n",
       "      <td>161674</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6</td>\n",
       "      <td>-1180.0</td>\n",
       "      <td>-1187.0</td>\n",
       "      <td>6948.360</td>\n",
       "      <td>6948.360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1330831</td>\n",
       "      <td>151639</td>\n",
       "      <td>0.0</td>\n",
       "      <td>34</td>\n",
       "      <td>-2156.0</td>\n",
       "      <td>-2156.0</td>\n",
       "      <td>1716.525</td>\n",
       "      <td>1716.525</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2085231</td>\n",
       "      <td>193053</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1</td>\n",
       "      <td>-63.0</td>\n",
       "      <td>-63.0</td>\n",
       "      <td>25425.000</td>\n",
       "      <td>25425.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2452527</td>\n",
       "      <td>199697</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>-2418.0</td>\n",
       "      <td>-2426.0</td>\n",
       "      <td>24350.130</td>\n",
       "      <td>24350.130</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2714724</td>\n",
       "      <td>167756</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>-1383.0</td>\n",
       "      <td>-1366.0</td>\n",
       "      <td>2165.040</td>\n",
       "      <td>2160.585</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_PREV  SK_ID_CURR  NUM_INSTALMENT_VERSION  NUM_INSTALMENT_NUMBER  \\\n",
       "0     1054186      161674                     1.0                      6   \n",
       "1     1330831      151639                     0.0                     34   \n",
       "2     2085231      193053                     2.0                      1   \n",
       "3     2452527      199697                     1.0                      3   \n",
       "4     2714724      167756                     1.0                      2   \n",
       "\n",
       "   DAYS_INSTALMENT  DAYS_ENTRY_PAYMENT  AMT_INSTALMENT  AMT_PAYMENT  \n",
       "0          -1180.0             -1187.0        6948.360     6948.360  \n",
       "1          -2156.0             -2156.0        1716.525     1716.525  \n",
       "2            -63.0               -63.0       25425.000    25425.000  \n",
       "3          -2418.0             -2426.0       24350.130    24350.130  \n",
       "4          -1383.0             -1366.0        2165.040     2160.585  "
      ]
     },
     "execution_count": 219,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "installments_payments.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 220,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>NUM_INSTALMENT_VERSION</th>\n",
       "      <th>NUM_INSTALMENT_NUMBER</th>\n",
       "      <th>DAYS_INSTALMENT</th>\n",
       "      <th>DAYS_ENTRY_PAYMENT</th>\n",
       "      <th>AMT_INSTALMENT</th>\n",
       "      <th>AMT_PAYMENT</th>\n",
       "      <th>DIFF_INST_PAYMENT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1054186</td>\n",
       "      <td>161674</td>\n",
       "      <td>1.0</td>\n",
       "      <td>6</td>\n",
       "      <td>-1180.0</td>\n",
       "      <td>-1187.0</td>\n",
       "      <td>6948.360</td>\n",
       "      <td>6948.360</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1330831</td>\n",
       "      <td>151639</td>\n",
       "      <td>0.0</td>\n",
       "      <td>34</td>\n",
       "      <td>-2156.0</td>\n",
       "      <td>-2156.0</td>\n",
       "      <td>1716.525</td>\n",
       "      <td>1716.525</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2085231</td>\n",
       "      <td>193053</td>\n",
       "      <td>2.0</td>\n",
       "      <td>1</td>\n",
       "      <td>-63.0</td>\n",
       "      <td>-63.0</td>\n",
       "      <td>25425.000</td>\n",
       "      <td>25425.000</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2452527</td>\n",
       "      <td>199697</td>\n",
       "      <td>1.0</td>\n",
       "      <td>3</td>\n",
       "      <td>-2418.0</td>\n",
       "      <td>-2426.0</td>\n",
       "      <td>24350.130</td>\n",
       "      <td>24350.130</td>\n",
       "      <td>0.000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2714724</td>\n",
       "      <td>167756</td>\n",
       "      <td>1.0</td>\n",
       "      <td>2</td>\n",
       "      <td>-1383.0</td>\n",
       "      <td>-1366.0</td>\n",
       "      <td>2165.040</td>\n",
       "      <td>2160.585</td>\n",
       "      <td>4.455</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_PREV  SK_ID_CURR  NUM_INSTALMENT_VERSION  NUM_INSTALMENT_NUMBER  \\\n",
       "0     1054186      161674                     1.0                      6   \n",
       "1     1330831      151639                     0.0                     34   \n",
       "2     2085231      193053                     2.0                      1   \n",
       "3     2452527      199697                     1.0                      3   \n",
       "4     2714724      167756                     1.0                      2   \n",
       "\n",
       "   DAYS_INSTALMENT  DAYS_ENTRY_PAYMENT  AMT_INSTALMENT  AMT_PAYMENT  \\\n",
       "0          -1180.0             -1187.0        6948.360     6948.360   \n",
       "1          -2156.0             -2156.0        1716.525     1716.525   \n",
       "2            -63.0               -63.0       25425.000    25425.000   \n",
       "3          -2418.0             -2426.0       24350.130    24350.130   \n",
       "4          -1383.0             -1366.0        2165.040     2160.585   \n",
       "\n",
       "   DIFF_INST_PAYMENT  \n",
       "0              0.000  \n",
       "1              0.000  \n",
       "2              0.000  \n",
       "3              0.000  \n",
       "4              4.455  "
      ]
     },
     "execution_count": 220,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "installments_payments['DIFF_INST_PAYMENT'] = (installments_payments['AMT_INSTALMENT'] - \n",
    "                                              installments_payments['AMT_PAYMENT'])\n",
    "installments_payments.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 221,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>AMT_INSTALMENT_MAX</th>\n",
       "      <th>AMT_INSTALMENT_MEAN</th>\n",
       "      <th>AMT_PAYMENT_MAX</th>\n",
       "      <th>AMT_PAYMENT_MEAN</th>\n",
       "      <th>DIFF_INST_PAYMENT_MAX</th>\n",
       "      <th>DIFF_INST_PAYMENT_MEAN</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>17397.900</td>\n",
       "      <td>5885.132143</td>\n",
       "      <td>17397.900</td>\n",
       "      <td>5885.132143</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>53093.745</td>\n",
       "      <td>11559.247105</td>\n",
       "      <td>53093.745</td>\n",
       "      <td>11559.247105</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>560835.360</td>\n",
       "      <td>64754.586000</td>\n",
       "      <td>560835.360</td>\n",
       "      <td>64754.586000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>10573.965</td>\n",
       "      <td>7096.155000</td>\n",
       "      <td>10573.965</td>\n",
       "      <td>7096.155000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>17656.245</td>\n",
       "      <td>6240.205000</td>\n",
       "      <td>17656.245</td>\n",
       "      <td>6240.205000</td>\n",
       "      <td>0.0</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  AMT_INSTALMENT_MAX  AMT_INSTALMENT_MEAN  AMT_PAYMENT_MAX  \\\n",
       "0      100001           17397.900          5885.132143        17397.900   \n",
       "1      100002           53093.745         11559.247105        53093.745   \n",
       "2      100003          560835.360         64754.586000       560835.360   \n",
       "3      100004           10573.965          7096.155000        10573.965   \n",
       "4      100005           17656.245          6240.205000        17656.245   \n",
       "\n",
       "   AMT_PAYMENT_MEAN  DIFF_INST_PAYMENT_MAX  DIFF_INST_PAYMENT_MEAN  \n",
       "0       5885.132143                    0.0                     0.0  \n",
       "1      11559.247105                    0.0                     0.0  \n",
       "2      64754.586000                    0.0                     0.0  \n",
       "3       7096.155000                    0.0                     0.0  \n",
       "4       6240.205000                    0.0                     0.0  "
      ]
     },
     "execution_count": 221,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_installments_payments = installments_payments.groupby('SK_ID_CURR').agg({'AMT_INSTALMENT':['max', 'mean'],\n",
    "                                                                   'AMT_PAYMENT':['max', 'mean'],\n",
    "                                                                   'DIFF_INST_PAYMENT':['max', 'mean']})\n",
    "agg_installments_payments.columns = agg_installments_payments.columns.map('_'.join).str.strip('_')\n",
    "agg_installments_payments.columns = map(str.upper, agg_installments_payments.columns)\n",
    "agg_installments_payments.reset_index(inplace=True)\n",
    "agg_installments_payments.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 222,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Feature scaling (min-max normalization)\n",
    "agg_installments_payments['AMT_INSTALMENT_MAX'] = ((agg_installments_payments['AMT_INSTALMENT_MAX'] - agg_installments_payments['AMT_INSTALMENT_MAX'].min())/(agg_installments_payments['AMT_INSTALMENT_MAX'].max() - agg_installments_payments['AMT_INSTALMENT_MAX'].min()))\n",
    "agg_installments_payments['AMT_INSTALMENT_MEAN'] = ((agg_installments_payments['AMT_INSTALMENT_MEAN'] - agg_installments_payments['AMT_INSTALMENT_MEAN'].min())/(agg_installments_payments['AMT_INSTALMENT_MEAN'].max() - agg_installments_payments['AMT_INSTALMENT_MEAN'].min()))\n",
    "agg_installments_payments['AMT_PAYMENT_MAX'] = ((agg_installments_payments['AMT_PAYMENT_MAX'] - agg_installments_payments['AMT_PAYMENT_MAX'].min())/(agg_installments_payments['AMT_PAYMENT_MAX'].max() - agg_installments_payments['AMT_PAYMENT_MAX'].min()))\n",
    "agg_installments_payments['AMT_PAYMENT_MEAN'] = ((agg_installments_payments['AMT_PAYMENT_MEAN'] - agg_installments_payments['AMT_PAYMENT_MEAN'].min())/(agg_installments_payments['AMT_PAYMENT_MEAN'].max() - agg_installments_payments['AMT_PAYMENT_MEAN'].min()))\n",
    "agg_installments_payments['DIFF_INST_PAYMENT_MAX'] = ((agg_installments_payments['DIFF_INST_PAYMENT_MAX'] - agg_installments_payments['DIFF_INST_PAYMENT_MAX'].min())/(agg_installments_payments['DIFF_INST_PAYMENT_MAX'].max() - agg_installments_payments['DIFF_INST_PAYMENT_MAX'].min()))\n",
    "agg_installments_payments['DIFF_INST_PAYMENT_MEAN'] = ((agg_installments_payments['DIFF_INST_PAYMENT_MEAN'] - agg_installments_payments['DIFF_INST_PAYMENT_MAX'].min())/(agg_installments_payments['DIFF_INST_PAYMENT_MEAN'].max() - agg_installments_payments['DIFF_INST_PAYMENT_MEAN'].min()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 223,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_processed_app_train = pd.merge(merged_processed_app_train, agg_installments_payments, how='left', on='SK_ID_CURR')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Debt to Income Ratio\n",
    "<p>Here I will calculate, using only the Credit Bureau information, an approximated debt to income ratio, that relates but not corresponds to the debt-to-income ratio (DTI) used in the mortgage industry.</p>\n",
    "<p> The total active debts divided by the applicant's income can be good estimator of his capacity to pay his debts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 224,
   "metadata": {},
   "outputs": [],
   "source": [
    "active_bureau = bureau[bureau['CREDIT_ACTIVE'] == 'Active']\n",
    "agg_active_bureau = active_bureau.groupby('SK_ID_CURR').agg({'AMT_CREDIT_SUM_DEBT':['sum']})\n",
    "agg_active_bureau.columns = agg_active_bureau.columns.map('_'.join).str.strip('_')\n",
    "agg_active_bureau.columns = map(str.upper, agg_active_bureau.columns)\n",
    "agg_active_bureau.reset_index(inplace=True)\n",
    "agg_active_bureau.rename(columns={'AMT_CREDIT_SUM_DEBT_SUM':'TOTAL_ACTIVE_DEBT'}, \n",
    "                 inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 225,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>TOTAL_ACTIVE_DEBT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>596686.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>245781.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>0.0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100005</td>\n",
       "      <td>568408.5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100008</td>\n",
       "      <td>240057.0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  TOTAL_ACTIVE_DEBT\n",
       "0      100001           596686.5\n",
       "1      100002           245781.0\n",
       "2      100003                0.0\n",
       "3      100005           568408.5\n",
       "4      100008           240057.0"
      ]
     },
     "execution_count": 225,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_active_bureau.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 226,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_processed_app_train = pd.merge(merged_processed_app_train, agg_active_bureau, how='left', on='SK_ID_CURR')\n",
    "\n",
    "merged_processed_app_train['TOTAL_ACTIVE_DEBT'].fillna(value=0.,inplace=True)\n",
    "\n",
    "merged_processed_app_train['DEBIT_TO_INCOME'] = (merged_processed_app_train['TOTAL_ACTIVE_DEBT'] / \n",
    "                                                 merged_processed_app_train['AMT_INCOME_TOTAL'])\n",
    "merged_processed_app_train = merged_processed_app_train.drop('TOTAL_ACTIVE_DEBT', axis=1)\n",
    "\n",
    "# Feature scaling (min-max normalization)\n",
    "merged_processed_app_train['DEBIT_TO_INCOME'] = ((merged_processed_app_train['DEBIT_TO_INCOME'] -\n",
    "                                                  merged_processed_app_train['DEBIT_TO_INCOME'].min())/\n",
    "                                                 (merged_processed_app_train['DEBIT_TO_INCOME'].max() - \n",
    "                                                  merged_processed_app_train['DEBIT_TO_INCOME'].min()))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 227,
   "metadata": {},
   "outputs": [],
   "source": [
    "# The remaining missing values will be replaced by the mean for each column\n",
    "merged_processed_app_train.fillna(value=merged_processed_app_train.mean(), inplace=True)\n",
    "merged_processed_app_train = merged_processed_app_train.drop('SK_ID_CURR', axis=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 228,
   "metadata": {},
   "outputs": [],
   "source": [
    "merged_processed_app_train.to_pickle('data/processed_data')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 229,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "\n",
      "Number of features: \t308\n",
      "Number of samples: \t302,054\n"
     ]
    }
   ],
   "source": [
    "print()\n",
    "print(\"Number of features: \\t{:,}\".format(len(merged_processed_app_train.columns)))\n",
    "print(\"Number of samples: \\t{:,}\".format(merged_processed_app_train.shape[0]))"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.1"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}