3_PredictingSportsWinnersTrees.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predicting Sports Winners with Decision Trees\n",
    "\n",
    "Here, we will look at predicting the winner of sports matches using a different type of classification algorithm: decision trees. These algorithms have a number of advantages over other algorithms. One of the main advantages is that they are readable by humans. In this way, decision trees can be used to learn a procedure,\n",
    "which could then be given to a human to perform if needed. Another advantage is that they work with a variety of features.\n",
    "\n",
    "## Loading the dataset\n",
    "\n",
    "We'll be looking at predicting the winner of games of the **National Basketball Association (NBA)**. Matches in the NBA are often close and can be decided in the last minute, making predicting the winner quite difficult. Many sports share this characteristic, whereby the expected winner could be beaten by another team on the right day.\n",
    "\n",
    "Various research into predicting the winner suggests that there may be an upper limit to sports outcome prediction accuracy which, depending on the sport, is between 70 percent and 80 percent accuracy. There is a significant amount of research being performed into sports prediction, often through data mining or statistics-based methods.\n",
    "\n",
    "## Collecting the data\n",
    "\n",
    "The data we will be using is the match history data for the NBA for the 2013-2014 season. The website http://Basketball-Reference.com contains a significant number of resources and statistics collected from the NBA and other leagues. To download the dataset, perform the following steps:\n",
    "\n",
    "1. Navigate to http://www.basketball-reference.com/leagues/NBA_2014_games.html in your web browser.\n",
    "2. Choose to get table as csv file for each month\n",
    "3. copy and paste the csv to your data folder and make a note of the path.\n",
    "\n",
    "This will download a **CSV** (short for **Comma Separated Values**) file containing the\n",
    "results of the 1,230 games in the regular season for the NBA.\n",
    "\n",
    "CSV files are simply text files where each line contains a new row and each value is separated by a comma (hence the name). CSV files can be created manually by simply typing into a text editor and saving with a *.csv* extension. They can also be opened in any program that can read text files, but can also be opened in Excel as a spreadsheet.\n",
    "\n",
    "We will load the file with the **pandas** (short for **Python Data Analysis**) library, which is an incredibly useful library for manipulating data. Python also contains a,built-in library called *csv* that supports reading and writing CSV files. However, we will use pandas, which provides more powerful functions that we will use later in the chapter for creating new features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [],
   "source": [
    "DATA = 'data/'\n",
    "NBA_2014_SEASON = DATA + 'leagues_NBA_2014_games_games.csv'\n",
    "NBA_EXTENDED_STANDINGS = DATA + 'leagues_NBA_2013_standings_expanded_standings.csv'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>Date</th>\n",
       "      <th>Start (ET)</th>\n",
       "      <th>Visitor/Neutral</th>\n",
       "      <th>PTS</th>\n",
       "      <th>Home/Neutral</th>\n",
       "      <th>PTS.1</th>\n",
       "      <th>Unnamed: 6</th>\n",
       "      <th>Unnamed: 7</th>\n",
       "      <th>Attend.</th>\n",
       "      <th>LOG</th>\n",
       "      <th>Arena</th>\n",
       "      <th>Notes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>Tue Oct 29 2013</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>87</td>\n",
       "      <td>Indiana Pacers</td>\n",
       "      <td>97</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18165</td>\n",
       "      <td>2:17</td>\n",
       "      <td>Bankers Life Fieldhouse</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>Tue Oct 29 2013</td>\n",
       "      <td>10:30p</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>103</td>\n",
       "      <td>Los Angeles Lakers</td>\n",
       "      <td>116</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18997</td>\n",
       "      <td>2:27</td>\n",
       "      <td>STAPLES Center</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>Tue Oct 29 2013</td>\n",
       "      <td>8:00p</td>\n",
       "      <td>Chicago Bulls</td>\n",
       "      <td>95</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>107</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19964</td>\n",
       "      <td>2:32</td>\n",
       "      <td>AmericanAirlines Arena</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>Wed Oct 30 2013</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>94</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>98</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20562</td>\n",
       "      <td>2:23</td>\n",
       "      <td>Quicken Loans Arena</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>Wed Oct 30 2013</td>\n",
       "      <td>8:30p</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>109</td>\n",
       "      <td>Dallas Mavericks</td>\n",
       "      <td>118</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19834</td>\n",
       "      <td>2:14</td>\n",
       "      <td>American Airlines Center</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "              Date Start (ET)       Visitor/Neutral  PTS         Home/Neutral  \\\n",
       "0  Tue Oct 29 2013      7:00p         Orlando Magic   87       Indiana Pacers   \n",
       "1  Tue Oct 29 2013     10:30p  Los Angeles Clippers  103   Los Angeles Lakers   \n",
       "2  Tue Oct 29 2013      8:00p         Chicago Bulls   95           Miami Heat   \n",
       "3  Wed Oct 30 2013      7:00p         Brooklyn Nets   94  Cleveland Cavaliers   \n",
       "4  Wed Oct 30 2013      8:30p         Atlanta Hawks  109     Dallas Mavericks   \n",
       "\n",
       "   PTS.1 Unnamed: 6 Unnamed: 7  Attend.   LOG                     Arena Notes  \n",
       "0     97  Box Score        NaN    18165  2:17   Bankers Life Fieldhouse   NaN  \n",
       "1    116  Box Score        NaN    18997  2:27            STAPLES Center   NaN  \n",
       "2    107  Box Score        NaN    19964  2:32    AmericanAirlines Arena   NaN  \n",
       "3     98  Box Score        NaN    20562  2:23       Quicken Loans Arena   NaN  \n",
       "4    118  Box Score        NaN    19834  2:14  American Airlines Center   NaN  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "\n",
    "# This is the entire season! Make sure to filter regular season\n",
    "# October 29, 2013 to April 16, 2014\n",
    "results = pd.read_csv(NBA_2014_SEASON)\n",
    "results.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Cleaning up the dataset\n",
    "\n",
    "After looking at the output, we can see a number of problems:\n",
    "- The date is just a string and not a date object\n",
    "- From visually inspecting the results, the headings aren't complete or correct\n",
    "\n",
    "These issues come from the data, and we could fix this by altering the data itself. However, in doing this, we could forget the steps we took or misapply them; that is, we can't replicate our results. As with the previous section where we used pipelines to track the transformations we made to a dataset, we will use pandas to apply\n",
    "transformations to the raw data itself."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "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>Date</th>\n",
       "      <th>Start (ET)</th>\n",
       "      <th>Visitor Team</th>\n",
       "      <th>VisitorPts</th>\n",
       "      <th>Home Team</th>\n",
       "      <th>HomePts</th>\n",
       "      <th>Score Type</th>\n",
       "      <th>OT?</th>\n",
       "      <th>Attend.</th>\n",
       "      <th>LOG</th>\n",
       "      <th>Arena</th>\n",
       "      <th>Notes</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>87</td>\n",
       "      <td>Indiana Pacers</td>\n",
       "      <td>97</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18165</td>\n",
       "      <td>2:17</td>\n",
       "      <td>Bankers Life Fieldhouse</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>10:30p</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>103</td>\n",
       "      <td>Los Angeles Lakers</td>\n",
       "      <td>116</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18997</td>\n",
       "      <td>2:27</td>\n",
       "      <td>STAPLES Center</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>8:00p</td>\n",
       "      <td>Chicago Bulls</td>\n",
       "      <td>95</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>107</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19964</td>\n",
       "      <td>2:32</td>\n",
       "      <td>AmericanAirlines Arena</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>94</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>98</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20562</td>\n",
       "      <td>2:23</td>\n",
       "      <td>Quicken Loans Arena</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>8:30p</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>109</td>\n",
       "      <td>Dallas Mavericks</td>\n",
       "      <td>118</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19834</td>\n",
       "      <td>2:14</td>\n",
       "      <td>American Airlines Center</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date Start (ET)          Visitor Team  VisitorPts  \\\n",
       "0 2013-10-29      7:00p         Orlando Magic          87   \n",
       "1 2013-10-29     10:30p  Los Angeles Clippers         103   \n",
       "2 2013-10-29      8:00p         Chicago Bulls          95   \n",
       "3 2013-10-30      7:00p         Brooklyn Nets          94   \n",
       "4 2013-10-30      8:30p         Atlanta Hawks         109   \n",
       "\n",
       "             Home Team  HomePts Score Type  OT?  Attend.   LOG  \\\n",
       "0       Indiana Pacers       97  Box Score  NaN    18165  2:17   \n",
       "1   Los Angeles Lakers      116  Box Score  NaN    18997  2:27   \n",
       "2           Miami Heat      107  Box Score  NaN    19964  2:32   \n",
       "3  Cleveland Cavaliers       98  Box Score  NaN    20562  2:23   \n",
       "4     Dallas Mavericks      118  Box Score  NaN    19834  2:14   \n",
       "\n",
       "                      Arena Notes  \n",
       "0   Bankers Life Fieldhouse   NaN  \n",
       "1            STAPLES Center   NaN  \n",
       "2    AmericanAirlines Arena   NaN  \n",
       "3       Quicken Loans Arena   NaN  \n",
       "4  American Airlines Center   NaN  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# Parse the date column as a date\n",
    "results = pd.read_csv(NBA_2014_SEASON, parse_dates=[\"Date\"])\n",
    "# Fix the name of the columns\n",
    "results.columns = ['Date', 'Start (ET)', 'Visitor Team', 'VisitorPts', 'Home Team', 'HomePts', 'Score Type', 'OT?', 'Attend.', 'LOG', 'Arena', 'Notes']\n",
    "results.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1319, 12)"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results.shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1230, 12)"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results = results[results['Date'] <= '2014-04-16']\n",
    "results.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now that we have our dataset, we can compute a **baseline**. A baseline is an accuracy that indicates an easy way to get a good accuracy. Any data mining solution should beat this.\n",
    "\n",
    "In each match, we have two teams: a home team and a visitor team. An obvious baseline, called the chance rate, is 50 percent. Choosing randomly will (over time) result in an accuracy of 50 percent.\n",
    "\n",
    "## Extracting new features\n",
    "\n",
    "We can now extract our features from this dataset by combining and comparing the existing data. First up, we need to specify our class value, which will give our classification algorithm something to compare against to see if its prediction is correct or not. This could be encoded in a number of ways; however, for this application, we will specify our class as 1 if the home team wins and 0 if the visitor team wins."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "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>Date</th>\n",
       "      <th>Start (ET)</th>\n",
       "      <th>Visitor Team</th>\n",
       "      <th>VisitorPts</th>\n",
       "      <th>Home Team</th>\n",
       "      <th>HomePts</th>\n",
       "      <th>Score Type</th>\n",
       "      <th>OT?</th>\n",
       "      <th>Attend.</th>\n",
       "      <th>LOG</th>\n",
       "      <th>Arena</th>\n",
       "      <th>Notes</th>\n",
       "      <th>HomeWin</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>87</td>\n",
       "      <td>Indiana Pacers</td>\n",
       "      <td>97</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18165</td>\n",
       "      <td>2:17</td>\n",
       "      <td>Bankers Life Fieldhouse</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>10:30p</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>103</td>\n",
       "      <td>Los Angeles Lakers</td>\n",
       "      <td>116</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18997</td>\n",
       "      <td>2:27</td>\n",
       "      <td>STAPLES Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>8:00p</td>\n",
       "      <td>Chicago Bulls</td>\n",
       "      <td>95</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>107</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19964</td>\n",
       "      <td>2:32</td>\n",
       "      <td>AmericanAirlines Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>94</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>98</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20562</td>\n",
       "      <td>2:23</td>\n",
       "      <td>Quicken Loans Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>8:30p</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>109</td>\n",
       "      <td>Dallas Mavericks</td>\n",
       "      <td>118</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19834</td>\n",
       "      <td>2:14</td>\n",
       "      <td>American Airlines Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date Start (ET)          Visitor Team  VisitorPts  \\\n",
       "0 2013-10-29      7:00p         Orlando Magic          87   \n",
       "1 2013-10-29     10:30p  Los Angeles Clippers         103   \n",
       "2 2013-10-29      8:00p         Chicago Bulls          95   \n",
       "3 2013-10-30      7:00p         Brooklyn Nets          94   \n",
       "4 2013-10-30      8:30p         Atlanta Hawks         109   \n",
       "\n",
       "             Home Team  HomePts Score Type  OT?  Attend.   LOG  \\\n",
       "0       Indiana Pacers       97  Box Score  NaN    18165  2:17   \n",
       "1   Los Angeles Lakers      116  Box Score  NaN    18997  2:27   \n",
       "2           Miami Heat      107  Box Score  NaN    19964  2:32   \n",
       "3  Cleveland Cavaliers       98  Box Score  NaN    20562  2:23   \n",
       "4     Dallas Mavericks      118  Box Score  NaN    19834  2:14   \n",
       "\n",
       "                      Arena Notes  HomeWin  \n",
       "0   Bankers Life Fieldhouse   NaN     True  \n",
       "1            STAPLES Center   NaN     True  \n",
       "2    AmericanAirlines Arena   NaN     True  \n",
       "3       Quicken Loans Arena   NaN     True  \n",
       "4  American Airlines Center   NaN     True  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "results[\"HomeWin\"] = results[\"VisitorPts\"] < results[\"HomePts\"]\n",
    "\n",
    "# Our \"class values\"\n",
    "y_true = results[\"HomeWin\"].values\n",
    "results.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Home Win percentage: 58.0%\n"
     ]
    }
   ],
   "source": [
    "home_win_percentage = 100 * results[\"HomeWin\"].sum() / results[\"HomeWin\"].count()\n",
    "print(f\"Home Win percentage: {home_win_percentage:.1f}%\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "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>Date</th>\n",
       "      <th>Start (ET)</th>\n",
       "      <th>Visitor Team</th>\n",
       "      <th>VisitorPts</th>\n",
       "      <th>Home Team</th>\n",
       "      <th>HomePts</th>\n",
       "      <th>Score Type</th>\n",
       "      <th>OT?</th>\n",
       "      <th>Attend.</th>\n",
       "      <th>LOG</th>\n",
       "      <th>Arena</th>\n",
       "      <th>Notes</th>\n",
       "      <th>HomeWin</th>\n",
       "      <th>HomeLastWin</th>\n",
       "      <th>VisitorLastWin</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>87</td>\n",
       "      <td>Indiana Pacers</td>\n",
       "      <td>97</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18165</td>\n",
       "      <td>2:17</td>\n",
       "      <td>Bankers Life Fieldhouse</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>10:30p</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>103</td>\n",
       "      <td>Los Angeles Lakers</td>\n",
       "      <td>116</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18997</td>\n",
       "      <td>2:27</td>\n",
       "      <td>STAPLES Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>8:00p</td>\n",
       "      <td>Chicago Bulls</td>\n",
       "      <td>95</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>107</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19964</td>\n",
       "      <td>2:32</td>\n",
       "      <td>AmericanAirlines Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>94</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>98</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20562</td>\n",
       "      <td>2:23</td>\n",
       "      <td>Quicken Loans Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>8:30p</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>109</td>\n",
       "      <td>Dallas Mavericks</td>\n",
       "      <td>118</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19834</td>\n",
       "      <td>2:14</td>\n",
       "      <td>American Airlines Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date Start (ET)          Visitor Team  VisitorPts  \\\n",
       "0 2013-10-29      7:00p         Orlando Magic          87   \n",
       "1 2013-10-29     10:30p  Los Angeles Clippers         103   \n",
       "2 2013-10-29      8:00p         Chicago Bulls          95   \n",
       "3 2013-10-30      7:00p         Brooklyn Nets          94   \n",
       "4 2013-10-30      8:30p         Atlanta Hawks         109   \n",
       "\n",
       "             Home Team  HomePts Score Type  OT?  Attend.   LOG  \\\n",
       "0       Indiana Pacers       97  Box Score  NaN    18165  2:17   \n",
       "1   Los Angeles Lakers      116  Box Score  NaN    18997  2:27   \n",
       "2           Miami Heat      107  Box Score  NaN    19964  2:32   \n",
       "3  Cleveland Cavaliers       98  Box Score  NaN    20562  2:23   \n",
       "4     Dallas Mavericks      118  Box Score  NaN    19834  2:14   \n",
       "\n",
       "                      Arena Notes  HomeWin  HomeLastWin  VisitorLastWin  \n",
       "0   Bankers Life Fieldhouse   NaN     True        False           False  \n",
       "1            STAPLES Center   NaN     True        False           False  \n",
       "2    AmericanAirlines Arena   NaN     True        False           False  \n",
       "3       Quicken Loans Arena   NaN     True        False           False  \n",
       "4  American Airlines Center   NaN     True        False           False  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "results[\"HomeLastWin\"] = False\n",
    "results[\"VisitorLastWin\"] = False\n",
    "# This creates two new columns, all set to False\n",
    "results.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The key of this dictionary will be the team and the value will be whether they won their previous game. We can then iterate over all the rows and update the current row with the team's last result."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "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>Date</th>\n",
       "      <th>Start (ET)</th>\n",
       "      <th>Visitor Team</th>\n",
       "      <th>VisitorPts</th>\n",
       "      <th>Home Team</th>\n",
       "      <th>HomePts</th>\n",
       "      <th>Score Type</th>\n",
       "      <th>OT?</th>\n",
       "      <th>Attend.</th>\n",
       "      <th>LOG</th>\n",
       "      <th>Arena</th>\n",
       "      <th>Notes</th>\n",
       "      <th>HomeWin</th>\n",
       "      <th>HomeLastWin</th>\n",
       "      <th>VisitorLastWin</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>2013-11-01</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>New Orleans Pelicans</td>\n",
       "      <td>90</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>110</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18846</td>\n",
       "      <td>2:13</td>\n",
       "      <td>Amway Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>2013-11-01</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Philadelphia 76ers</td>\n",
       "      <td>109</td>\n",
       "      <td>Washington Wizards</td>\n",
       "      <td>102</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17160</td>\n",
       "      <td>2:06</td>\n",
       "      <td>Verizon Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>2013-11-01</td>\n",
       "      <td>7:30p</td>\n",
       "      <td>Toronto Raptors</td>\n",
       "      <td>95</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>102</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18118</td>\n",
       "      <td>2:12</td>\n",
       "      <td>Philips Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>True</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>2013-11-01</td>\n",
       "      <td>7:30p</td>\n",
       "      <td>Milwaukee Bucks</td>\n",
       "      <td>105</td>\n",
       "      <td>Boston Celtics</td>\n",
       "      <td>98</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18624</td>\n",
       "      <td>2:23</td>\n",
       "      <td>TD Garden</td>\n",
       "      <td>NaN</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>2013-11-01</td>\n",
       "      <td>8:00p</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>100</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>101</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>17732</td>\n",
       "      <td>2:54</td>\n",
       "      <td>Barclays Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>False</td>\n",
       "      <td>False</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         Date Start (ET)          Visitor Team  VisitorPts  \\\n",
       "20 2013-11-01      7:00p  New Orleans Pelicans          90   \n",
       "21 2013-11-01      7:00p    Philadelphia 76ers         109   \n",
       "22 2013-11-01      7:30p       Toronto Raptors          95   \n",
       "23 2013-11-01      7:30p       Milwaukee Bucks         105   \n",
       "24 2013-11-01      8:00p            Miami Heat         100   \n",
       "\n",
       "             Home Team  HomePts Score Type  OT?  Attend.   LOG  \\\n",
       "20       Orlando Magic      110  Box Score  NaN    18846  2:13   \n",
       "21  Washington Wizards      102  Box Score  NaN    17160  2:06   \n",
       "22       Atlanta Hawks      102  Box Score  NaN    18118  2:12   \n",
       "23      Boston Celtics       98  Box Score  NaN    18624  2:23   \n",
       "24       Brooklyn Nets      101  Box Score  NaN    17732  2:54   \n",
       "\n",
       "              Arena Notes  HomeWin HomeLastWin VisitorLastWin  \n",
       "20     Amway Center   NaN     True       False          False  \n",
       "21   Verizon Center   NaN    False       False           True  \n",
       "22    Philips Arena   NaN     True       False           True  \n",
       "23        TD Garden   NaN    False       False          False  \n",
       "24  Barclays Center   NaN     True       False          False  "
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Now compute the actual values for these\n",
    "# Did the home and visitor teams win their last game?\n",
    "from collections import defaultdict\n",
    "\n",
    "\n",
    "won_last = defaultdict(int)\n",
    "for index, row in results.iterrows():  # Note that this is not efficient\n",
    "    home_team = row[\"Home Team\"]\n",
    "    visitor_team = row[\"Visitor Team\"]\n",
    "    row[\"HomeLastWin\"] = won_last[home_team]\n",
    "    row[\"VisitorLastWin\"] = won_last[visitor_team]\n",
    "    results.iloc[index] = row    \n",
    "    # Set current win\n",
    "    won_last[home_team] = row[\"HomeWin\"]\n",
    "    won_last[visitor_team] = not row[\"HomeWin\"]\n",
    "results.iloc[20:25]"
   ]
  },
  {
   "attachments": {
    "image.png": {
     "image/png": "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"
    }
   },
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Decision trees\n",
    "\n",
    "Decision trees are a class of supervised learning algorithm like a flow chart that consists of a sequence of nodes, where the values for a sample are used to make a decision on the next node to go to.\n",
    "\n",
    "![image.png](attachment:image.png)\n",
    "\n",
    "As with most classification algorithms, there are two components:\n",
    "\n",
    "- The first is the training stage, where a tree is built using training data. While the nearest neighbor algorithm from the previous chapter did not have a training phase, it is needed for decision trees. In this way, the nearest neighbor algorithm is a *lazy learner*, only doing any work when it needs to make a prediction. In contrast, decision trees, like most classification methods, are *eager learners*, undertaking work at the training stage.\n",
    "\n",
    "• The second is the predicting stage, where the trained tree is used to predict the classification of new samples. Using the previous example tree, a data point of [\"is raining\", \"very windy\"] would be classed\n",
    "as \"bad weather\".\n",
    "\n",
    "The `scikit-learn` package implements the **CART** (**Classification and Regression Trees**) algorithm as its default decision tree class, which can use both categorical and continuous features.\n",
    "\n",
    "## Parameters in decision trees\n",
    "\n",
    "One of the most important features for a decision tree is the stopping criterion. Using such specific nodes\n",
    "can results in trees that significantly overfit the training data. Instead, a stopping criterion can be used to ensure that the decision tree does not reach this exactness.\n",
    "\n",
    "Instead of using a stopping criterion, the tree could be created in full and then trimmed. This trimming process removes nodes that do not provide much information to the overall process. This is known as pruning.\n",
    "\n",
    "The decision tree implementation in scikit-learn provides a method to stop the\n",
    "building of a tree using the following options:\n",
    "\n",
    "- `min_samples_split`: This specifies how many samples are needed in order to create a new node in the decision tree\n",
    "- `min_samples_leaf`: This specifies how many samples must be resulting from a node for it to stay\n",
    "\n",
    "Another parameter for decision tress is the criterion for creating a decision. Gini impurity and Information gain are two popular ones:\n",
    "\n",
    "- **Gini impurity**: This is a measure of how often a decision node would incorrectly predict a sample's class\n",
    "- **Information gain**: This uses information-theory-based entropy to indicate how much extra information is gained by the decision node\n",
    "\n",
    "## Using decision trees\n",
    "\n",
    "We can import the DecisionTreeClassifier class and create a decision tree using scikit-learn:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "clf = DecisionTreeClassifier(random_state=14)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using just the last result from the home and visitor teams\n",
      "Accuracy: 58.0%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import cross_val_score\n",
    "\n",
    "# Create a dataset with just the neccessary information\n",
    "X_previouswins = results[[\"HomeLastWin\", \"VisitorLastWin\"]].values\n",
    "clf = DecisionTreeClassifier(random_state=14)\n",
    "scores = cross_val_score(clf, X_previouswins, y_true, scoring='accuracy')\n",
    "print(\"Using just the last result from the home and visitor teams\")\n",
    "print(f\"Accuracy: {np.mean(scores) * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This scores 58 percent: we are better than choosing randomly! We should be able to do better. Feature engineering is one of the most difficult tasks in data mining, and choosing good features is key to getting good outcomes—more so than choosing the right algorithm!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# What about win streaks?\n",
    "results[\"HomeWinStreak\"] = 0\n",
    "results[\"VisitorWinStreak\"] = 0\n",
    "# Did the home and visitor teams win their last game?\n",
    "from collections import defaultdict\n",
    "win_streak = defaultdict(int)\n",
    "\n",
    "for index, row in results.iterrows():  # Note that this is not efficient\n",
    "    home_team = row[\"Home Team\"]\n",
    "    visitor_team = row[\"Visitor Team\"]\n",
    "    row[\"HomeWinStreak\"] = win_streak[home_team]\n",
    "    row[\"VisitorWinStreak\"] = win_streak[visitor_team]\n",
    "    results.iloc[index] = row    \n",
    "    # Set current win\n",
    "    if row[\"HomeWin\"]:\n",
    "        win_streak[home_team] += 1\n",
    "        win_streak[visitor_team] = 0\n",
    "    else:\n",
    "        win_streak[home_team] = 0\n",
    "        win_streak[visitor_team] += 1\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using whether the home team is ranked higher\n",
      "Accuracy: 58.3%\n"
     ]
    }
   ],
   "source": [
    "clf = DecisionTreeClassifier(random_state=14)\n",
    "X_winstreak =  results[[\"HomeLastWin\", \"VisitorLastWin\", \"HomeWinStreak\", \"VisitorWinStreak\"]].values\n",
    "scores = cross_val_score(clf, X_winstreak, y_true, scoring='accuracy')\n",
    "print(\"Using whether the home team is ranked higher\")\n",
    "print(f\"Accuracy: {np.mean(scores) * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Glossary for expanded standings\n",
    "\n",
    "Also view explanations by holding mouse over column headers\n",
    "- Rk -- Rank\n",
    "- Overall -- Overall W-L record\n",
    "- Place\n",
    "- Home -- W-L record in home games\n",
    "- Road -- W-L record in road games\n",
    "- Conference\n",
    "- E -- W-L record versus Eastern Conference teams\n",
    "- W -- W-L record versus Western Conference teams\n",
    "- Division\n",
    "- A -- W-L record versus Atlantic Division teams\n",
    "- C -- W-L record versus Central Division teams\n",
    "- SE -- W-L record versus Southeast Division teams\n",
    "- NW -- W-L record versus Northwest Division teams\n",
    "- P -- W-L record versus Pacific Division teams\n",
    "- SW -- W-L record versus Southwest Division teams\n",
    "- All-Star\n",
    "- Pre -- W-L records before the All-Star Game\n",
    "- Post -- W-L records after the All-Star Game\n",
    "- Margin\n",
    "- ≤3 -- W-L record in games decided by 3 or fewer points\n",
    "- ≥10 -- W-L record in games decided by 10 or more points\n",
    "- Month\n",
    "- Oct -- W-L record in the month of October\n",
    "- Nov -- W-L record in the month of November\n",
    "- Dec -- W-L record in the month of December\n",
    "- Jan -- W-L record in the month of January\n",
    "- Feb -- W-L record in the month of February\n",
    "- Mar -- W-L record in the month of March\n",
    "- Apr -- W-L record in the month of April\n",
    "- Rk -- Rank\n",
    "- Overall -- Overall W-L record\n",
    "- Place\n",
    "- Home -- W-L record in home games\n",
    "- Road -- W-L record in road games\n",
    "- Conference\n",
    "- E -- W-L record versus Eastern Conference teams\n",
    "- W -- W-L record versus Western Conference teams\n",
    "- Division\n",
    "- A -- W-L record versus Atlantic Division teams\n",
    "- C -- W-L record versus Central Division teams\n",
    "- SE -- W-L record versus Southeast Division teams\n",
    "- NW -- W-L record versus Northwest Division teams\n",
    "- P -- W-L record versus Pacific Division teams\n",
    "- SW -- W-L record versus Southwest Division teams\n",
    "- All-Star\n",
    "- Pre -- W-L records before the All-Star Game\n",
    "- Post -- W-L records after the All-Star Game\n",
    "- Margin\n",
    "- ≤3 -- W-L record in games decided by 3 or fewer points\n",
    "- ≥10 -- W-L record in games decided by 10 or more points\n",
    "- Month\n",
    "- Oct -- W-L record in the month of October\n",
    "- Nov -- W-L record in the month of November\n",
    "- Dec -- W-L record in the month of December\n",
    "- Jan -- W-L record in the month of January\n",
    "- Feb -- W-L record in the month of February\n",
    "- Mar -- W-L record in the month of March\n",
    "- Apr -- W-L record in the month of April"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "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>Rk</th>\n",
       "      <th>Team</th>\n",
       "      <th>Overall</th>\n",
       "      <th>Home</th>\n",
       "      <th>Road</th>\n",
       "      <th>E</th>\n",
       "      <th>W</th>\n",
       "      <th>A</th>\n",
       "      <th>C</th>\n",
       "      <th>SE</th>\n",
       "      <th>...</th>\n",
       "      <th>Post</th>\n",
       "      <th>≤3</th>\n",
       "      <th>≥10</th>\n",
       "      <th>Oct</th>\n",
       "      <th>Nov</th>\n",
       "      <th>Dec</th>\n",
       "      <th>Jan</th>\n",
       "      <th>Feb</th>\n",
       "      <th>Mar</th>\n",
       "      <th>Apr</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>San Antonio Spurs</td>\n",
       "      <td>62-20</td>\n",
       "      <td>32-9</td>\n",
       "      <td>30-11</td>\n",
       "      <td>24-6</td>\n",
       "      <td>38-14</td>\n",
       "      <td>8-2</td>\n",
       "      <td>7-3</td>\n",
       "      <td>9-1</td>\n",
       "      <td>...</td>\n",
       "      <td>24-5</td>\n",
       "      <td>7-0</td>\n",
       "      <td>35-12</td>\n",
       "      <td>1-0</td>\n",
       "      <td>13-3</td>\n",
       "      <td>11-4</td>\n",
       "      <td>8-6</td>\n",
       "      <td>9-3</td>\n",
       "      <td>16-0</td>\n",
       "      <td>4-4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>Oklahoma City Thunder</td>\n",
       "      <td>59-23</td>\n",
       "      <td>34-7</td>\n",
       "      <td>25-16</td>\n",
       "      <td>23-7</td>\n",
       "      <td>36-16</td>\n",
       "      <td>8-2</td>\n",
       "      <td>8-2</td>\n",
       "      <td>7-3</td>\n",
       "      <td>...</td>\n",
       "      <td>16-11</td>\n",
       "      <td>10-4</td>\n",
       "      <td>33-8</td>\n",
       "      <td>1-0</td>\n",
       "      <td>10-3</td>\n",
       "      <td>14-3</td>\n",
       "      <td>13-4</td>\n",
       "      <td>6-5</td>\n",
       "      <td>10-4</td>\n",
       "      <td>5-4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>57-25</td>\n",
       "      <td>34-7</td>\n",
       "      <td>23-18</td>\n",
       "      <td>21-9</td>\n",
       "      <td>36-16</td>\n",
       "      <td>9-1</td>\n",
       "      <td>7-3</td>\n",
       "      <td>5-5</td>\n",
       "      <td>...</td>\n",
       "      <td>20-7</td>\n",
       "      <td>4-3</td>\n",
       "      <td>30-9</td>\n",
       "      <td>1-1</td>\n",
       "      <td>11-4</td>\n",
       "      <td>9-7</td>\n",
       "      <td>12-4</td>\n",
       "      <td>7-4</td>\n",
       "      <td>13-2</td>\n",
       "      <td>4-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>Indiana Pacers</td>\n",
       "      <td>56-26</td>\n",
       "      <td>35-6</td>\n",
       "      <td>21-20</td>\n",
       "      <td>38-14</td>\n",
       "      <td>18-12</td>\n",
       "      <td>15-3</td>\n",
       "      <td>12-4</td>\n",
       "      <td>11-7</td>\n",
       "      <td>...</td>\n",
       "      <td>16-14</td>\n",
       "      <td>4-3</td>\n",
       "      <td>29-17</td>\n",
       "      <td>2-0</td>\n",
       "      <td>13-1</td>\n",
       "      <td>10-4</td>\n",
       "      <td>10-5</td>\n",
       "      <td>9-3</td>\n",
       "      <td>8-10</td>\n",
       "      <td>4-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>Houston Rockets</td>\n",
       "      <td>54-28</td>\n",
       "      <td>33-8</td>\n",
       "      <td>21-20</td>\n",
       "      <td>23-7</td>\n",
       "      <td>31-21</td>\n",
       "      <td>7-3</td>\n",
       "      <td>8-2</td>\n",
       "      <td>8-2</td>\n",
       "      <td>...</td>\n",
       "      <td>18-11</td>\n",
       "      <td>7-6</td>\n",
       "      <td>30-9</td>\n",
       "      <td>1-0</td>\n",
       "      <td>12-5</td>\n",
       "      <td>8-8</td>\n",
       "      <td>10-4</td>\n",
       "      <td>8-2</td>\n",
       "      <td>10-4</td>\n",
       "      <td>5-5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>54-28</td>\n",
       "      <td>32-9</td>\n",
       "      <td>22-19</td>\n",
       "      <td>34-18</td>\n",
       "      <td>20-10</td>\n",
       "      <td>9-9</td>\n",
       "      <td>13-5</td>\n",
       "      <td>12-4</td>\n",
       "      <td>...</td>\n",
       "      <td>17-14</td>\n",
       "      <td>8-7</td>\n",
       "      <td>29-10</td>\n",
       "      <td>1-1</td>\n",
       "      <td>12-2</td>\n",
       "      <td>11-4</td>\n",
       "      <td>8-6</td>\n",
       "      <td>9-1</td>\n",
       "      <td>10-8</td>\n",
       "      <td>3-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>Portland Trail Blazers</td>\n",
       "      <td>54-28</td>\n",
       "      <td>31-10</td>\n",
       "      <td>23-18</td>\n",
       "      <td>23-7</td>\n",
       "      <td>31-21</td>\n",
       "      <td>9-1</td>\n",
       "      <td>9-1</td>\n",
       "      <td>5-5</td>\n",
       "      <td>...</td>\n",
       "      <td>18-11</td>\n",
       "      <td>8-9</td>\n",
       "      <td>24-13</td>\n",
       "      <td>0-1</td>\n",
       "      <td>13-2</td>\n",
       "      <td>12-4</td>\n",
       "      <td>8-6</td>\n",
       "      <td>7-5</td>\n",
       "      <td>8-9</td>\n",
       "      <td>6-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>Golden State Warriors</td>\n",
       "      <td>51-31</td>\n",
       "      <td>27-14</td>\n",
       "      <td>24-17</td>\n",
       "      <td>20-10</td>\n",
       "      <td>31-21</td>\n",
       "      <td>7-3</td>\n",
       "      <td>7-3</td>\n",
       "      <td>6-4</td>\n",
       "      <td>...</td>\n",
       "      <td>20-9</td>\n",
       "      <td>11-8</td>\n",
       "      <td>28-9</td>\n",
       "      <td>1-1</td>\n",
       "      <td>8-7</td>\n",
       "      <td>11-5</td>\n",
       "      <td>9-6</td>\n",
       "      <td>7-4</td>\n",
       "      <td>9-5</td>\n",
       "      <td>6-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>Memphis Grizzlies</td>\n",
       "      <td>50-32</td>\n",
       "      <td>27-14</td>\n",
       "      <td>23-18</td>\n",
       "      <td>21-9</td>\n",
       "      <td>29-23</td>\n",
       "      <td>6-4</td>\n",
       "      <td>7-3</td>\n",
       "      <td>8-2</td>\n",
       "      <td>...</td>\n",
       "      <td>21-9</td>\n",
       "      <td>10-4</td>\n",
       "      <td>18-13</td>\n",
       "      <td>0-1</td>\n",
       "      <td>8-7</td>\n",
       "      <td>5-9</td>\n",
       "      <td>12-3</td>\n",
       "      <td>7-5</td>\n",
       "      <td>12-5</td>\n",
       "      <td>6-2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>Dallas Mavericks</td>\n",
       "      <td>49-33</td>\n",
       "      <td>26-15</td>\n",
       "      <td>23-18</td>\n",
       "      <td>20-10</td>\n",
       "      <td>29-23</td>\n",
       "      <td>5-5</td>\n",
       "      <td>9-1</td>\n",
       "      <td>6-4</td>\n",
       "      <td>...</td>\n",
       "      <td>17-11</td>\n",
       "      <td>11-11</td>\n",
       "      <td>23-10</td>\n",
       "      <td>1-0</td>\n",
       "      <td>9-8</td>\n",
       "      <td>8-5</td>\n",
       "      <td>9-8</td>\n",
       "      <td>9-3</td>\n",
       "      <td>8-6</td>\n",
       "      <td>5-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>11</td>\n",
       "      <td>Chicago Bulls</td>\n",
       "      <td>48-34</td>\n",
       "      <td>27-14</td>\n",
       "      <td>21-20</td>\n",
       "      <td>35-17</td>\n",
       "      <td>13-17</td>\n",
       "      <td>12-6</td>\n",
       "      <td>11-5</td>\n",
       "      <td>12-6</td>\n",
       "      <td>...</td>\n",
       "      <td>21-9</td>\n",
       "      <td>6-5</td>\n",
       "      <td>27-19</td>\n",
       "      <td>1-1</td>\n",
       "      <td>6-7</td>\n",
       "      <td>5-10</td>\n",
       "      <td>11-4</td>\n",
       "      <td>9-4</td>\n",
       "      <td>10-6</td>\n",
       "      <td>6-2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>12</td>\n",
       "      <td>Toronto Raptors</td>\n",
       "      <td>48-34</td>\n",
       "      <td>26-15</td>\n",
       "      <td>22-19</td>\n",
       "      <td>32-20</td>\n",
       "      <td>16-14</td>\n",
       "      <td>11-5</td>\n",
       "      <td>13-5</td>\n",
       "      <td>8-10</td>\n",
       "      <td>...</td>\n",
       "      <td>20-10</td>\n",
       "      <td>3-8</td>\n",
       "      <td>24-6</td>\n",
       "      <td>1-0</td>\n",
       "      <td>5-9</td>\n",
       "      <td>8-6</td>\n",
       "      <td>11-6</td>\n",
       "      <td>7-5</td>\n",
       "      <td>10-6</td>\n",
       "      <td>6-2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>13</td>\n",
       "      <td>Phoenix Suns</td>\n",
       "      <td>48-34</td>\n",
       "      <td>26-15</td>\n",
       "      <td>22-19</td>\n",
       "      <td>20-10</td>\n",
       "      <td>28-24</td>\n",
       "      <td>7-3</td>\n",
       "      <td>6-4</td>\n",
       "      <td>7-3</td>\n",
       "      <td>...</td>\n",
       "      <td>18-13</td>\n",
       "      <td>2-9</td>\n",
       "      <td>19-7</td>\n",
       "      <td>1-0</td>\n",
       "      <td>8-8</td>\n",
       "      <td>10-3</td>\n",
       "      <td>9-7</td>\n",
       "      <td>6-6</td>\n",
       "      <td>10-6</td>\n",
       "      <td>4-4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>14</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>44-38</td>\n",
       "      <td>28-13</td>\n",
       "      <td>16-25</td>\n",
       "      <td>26-26</td>\n",
       "      <td>18-12</td>\n",
       "      <td>9-7</td>\n",
       "      <td>7-11</td>\n",
       "      <td>10-8</td>\n",
       "      <td>...</td>\n",
       "      <td>20-11</td>\n",
       "      <td>8-3</td>\n",
       "      <td>17-20</td>\n",
       "      <td>0-1</td>\n",
       "      <td>5-11</td>\n",
       "      <td>5-9</td>\n",
       "      <td>10-3</td>\n",
       "      <td>7-5</td>\n",
       "      <td>12-4</td>\n",
       "      <td>5-5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>15</td>\n",
       "      <td>Washington Wizards</td>\n",
       "      <td>44-38</td>\n",
       "      <td>22-19</td>\n",
       "      <td>22-19</td>\n",
       "      <td>33-19</td>\n",
       "      <td>11-19</td>\n",
       "      <td>13-5</td>\n",
       "      <td>10-8</td>\n",
       "      <td>10-6</td>\n",
       "      <td>...</td>\n",
       "      <td>19-11</td>\n",
       "      <td>5-9</td>\n",
       "      <td>19-14</td>\n",
       "      <td>0-1</td>\n",
       "      <td>8-8</td>\n",
       "      <td>6-5</td>\n",
       "      <td>8-9</td>\n",
       "      <td>8-5</td>\n",
       "      <td>8-8</td>\n",
       "      <td>6-2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>16</td>\n",
       "      <td>Charlotte Bobcats</td>\n",
       "      <td>43-39</td>\n",
       "      <td>25-16</td>\n",
       "      <td>18-23</td>\n",
       "      <td>30-22</td>\n",
       "      <td>13-17</td>\n",
       "      <td>11-7</td>\n",
       "      <td>13-5</td>\n",
       "      <td>6-10</td>\n",
       "      <td>...</td>\n",
       "      <td>20-9</td>\n",
       "      <td>8-8</td>\n",
       "      <td>18-19</td>\n",
       "      <td>0-1</td>\n",
       "      <td>8-8</td>\n",
       "      <td>6-9</td>\n",
       "      <td>7-9</td>\n",
       "      <td>6-4</td>\n",
       "      <td>9-7</td>\n",
       "      <td>7-1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>17</td>\n",
       "      <td>Minnesota Timberwolves</td>\n",
       "      <td>40-42</td>\n",
       "      <td>24-17</td>\n",
       "      <td>16-25</td>\n",
       "      <td>17-13</td>\n",
       "      <td>23-29</td>\n",
       "      <td>5-5</td>\n",
       "      <td>7-3</td>\n",
       "      <td>5-5</td>\n",
       "      <td>...</td>\n",
       "      <td>15-14</td>\n",
       "      <td>5-6</td>\n",
       "      <td>26-18</td>\n",
       "      <td>1-0</td>\n",
       "      <td>8-9</td>\n",
       "      <td>6-7</td>\n",
       "      <td>8-7</td>\n",
       "      <td>5-6</td>\n",
       "      <td>8-8</td>\n",
       "      <td>4-5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>18</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>38-44</td>\n",
       "      <td>24-17</td>\n",
       "      <td>14-27</td>\n",
       "      <td>28-24</td>\n",
       "      <td>10-20</td>\n",
       "      <td>10-8</td>\n",
       "      <td>10-8</td>\n",
       "      <td>8-8</td>\n",
       "      <td>...</td>\n",
       "      <td>13-18</td>\n",
       "      <td>8-9</td>\n",
       "      <td>15-16</td>\n",
       "      <td>0-1</td>\n",
       "      <td>9-8</td>\n",
       "      <td>9-5</td>\n",
       "      <td>6-7</td>\n",
       "      <td>2-10</td>\n",
       "      <td>6-10</td>\n",
       "      <td>6-3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>19</td>\n",
       "      <td>New York Knicks</td>\n",
       "      <td>37-45</td>\n",
       "      <td>19-22</td>\n",
       "      <td>18-23</td>\n",
       "      <td>26-26</td>\n",
       "      <td>11-19</td>\n",
       "      <td>10-6</td>\n",
       "      <td>9-9</td>\n",
       "      <td>7-11</td>\n",
       "      <td>...</td>\n",
       "      <td>17-13</td>\n",
       "      <td>3-8</td>\n",
       "      <td>19-22</td>\n",
       "      <td>1-1</td>\n",
       "      <td>2-11</td>\n",
       "      <td>6-9</td>\n",
       "      <td>10-6</td>\n",
       "      <td>2-11</td>\n",
       "      <td>11-5</td>\n",
       "      <td>5-2</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>20</td>\n",
       "      <td>Denver Nuggets</td>\n",
       "      <td>36-46</td>\n",
       "      <td>22-19</td>\n",
       "      <td>14-27</td>\n",
       "      <td>16-14</td>\n",
       "      <td>20-32</td>\n",
       "      <td>5-5</td>\n",
       "      <td>5-5</td>\n",
       "      <td>6-4</td>\n",
       "      <td>...</td>\n",
       "      <td>12-19</td>\n",
       "      <td>8-5</td>\n",
       "      <td>16-24</td>\n",
       "      <td>0-1</td>\n",
       "      <td>9-5</td>\n",
       "      <td>5-10</td>\n",
       "      <td>8-7</td>\n",
       "      <td>3-9</td>\n",
       "      <td>7-10</td>\n",
       "      <td>4-4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>21</td>\n",
       "      <td>New Orleans Pelicans</td>\n",
       "      <td>34-48</td>\n",
       "      <td>22-19</td>\n",
       "      <td>12-29</td>\n",
       "      <td>19-11</td>\n",
       "      <td>15-37</td>\n",
       "      <td>6-4</td>\n",
       "      <td>8-2</td>\n",
       "      <td>5-5</td>\n",
       "      <td>...</td>\n",
       "      <td>11-19</td>\n",
       "      <td>7-8</td>\n",
       "      <td>10-21</td>\n",
       "      <td>0-1</td>\n",
       "      <td>7-7</td>\n",
       "      <td>7-7</td>\n",
       "      <td>5-11</td>\n",
       "      <td>4-9</td>\n",
       "      <td>9-7</td>\n",
       "      <td>2-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>22</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>33-49</td>\n",
       "      <td>19-22</td>\n",
       "      <td>14-27</td>\n",
       "      <td>21-31</td>\n",
       "      <td>12-18</td>\n",
       "      <td>8-10</td>\n",
       "      <td>7-9</td>\n",
       "      <td>6-12</td>\n",
       "      <td>...</td>\n",
       "      <td>13-16</td>\n",
       "      <td>7-5</td>\n",
       "      <td>13-27</td>\n",
       "      <td>1-0</td>\n",
       "      <td>4-12</td>\n",
       "      <td>5-9</td>\n",
       "      <td>6-9</td>\n",
       "      <td>8-6</td>\n",
       "      <td>6-9</td>\n",
       "      <td>3-4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>23</td>\n",
       "      <td>Detroit Pistons</td>\n",
       "      <td>29-53</td>\n",
       "      <td>17-24</td>\n",
       "      <td>12-29</td>\n",
       "      <td>23-29</td>\n",
       "      <td>6-24</td>\n",
       "      <td>11-7</td>\n",
       "      <td>6-10</td>\n",
       "      <td>6-12</td>\n",
       "      <td>...</td>\n",
       "      <td>7-23</td>\n",
       "      <td>2-6</td>\n",
       "      <td>16-25</td>\n",
       "      <td>1-0</td>\n",
       "      <td>5-10</td>\n",
       "      <td>8-9</td>\n",
       "      <td>4-8</td>\n",
       "      <td>5-8</td>\n",
       "      <td>4-12</td>\n",
       "      <td>2-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>24</td>\n",
       "      <td>Sacramento Kings</td>\n",
       "      <td>28-54</td>\n",
       "      <td>17-24</td>\n",
       "      <td>11-30</td>\n",
       "      <td>13-17</td>\n",
       "      <td>15-37</td>\n",
       "      <td>5-5</td>\n",
       "      <td>4-6</td>\n",
       "      <td>4-6</td>\n",
       "      <td>...</td>\n",
       "      <td>10-19</td>\n",
       "      <td>4-9</td>\n",
       "      <td>12-26</td>\n",
       "      <td>1-0</td>\n",
       "      <td>3-10</td>\n",
       "      <td>6-10</td>\n",
       "      <td>5-11</td>\n",
       "      <td>5-7</td>\n",
       "      <td>6-10</td>\n",
       "      <td>2-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>25</td>\n",
       "      <td>Los Angeles Lakers</td>\n",
       "      <td>27-55</td>\n",
       "      <td>14-27</td>\n",
       "      <td>13-28</td>\n",
       "      <td>12-18</td>\n",
       "      <td>15-37</td>\n",
       "      <td>6-4</td>\n",
       "      <td>3-7</td>\n",
       "      <td>3-7</td>\n",
       "      <td>...</td>\n",
       "      <td>9-20</td>\n",
       "      <td>5-5</td>\n",
       "      <td>12-33</td>\n",
       "      <td>1-1</td>\n",
       "      <td>8-7</td>\n",
       "      <td>4-11</td>\n",
       "      <td>3-12</td>\n",
       "      <td>4-8</td>\n",
       "      <td>5-9</td>\n",
       "      <td>2-7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>26</td>\n",
       "      <td>Boston Celtics</td>\n",
       "      <td>25-57</td>\n",
       "      <td>16-25</td>\n",
       "      <td>9-32</td>\n",
       "      <td>21-31</td>\n",
       "      <td>4-26</td>\n",
       "      <td>5-11</td>\n",
       "      <td>6-12</td>\n",
       "      <td>10-8</td>\n",
       "      <td>...</td>\n",
       "      <td>6-22</td>\n",
       "      <td>5-9</td>\n",
       "      <td>9-23</td>\n",
       "      <td>0-1</td>\n",
       "      <td>7-11</td>\n",
       "      <td>6-6</td>\n",
       "      <td>2-15</td>\n",
       "      <td>5-6</td>\n",
       "      <td>3-12</td>\n",
       "      <td>2-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>27</td>\n",
       "      <td>Utah Jazz</td>\n",
       "      <td>25-57</td>\n",
       "      <td>16-25</td>\n",
       "      <td>9-32</td>\n",
       "      <td>12-18</td>\n",
       "      <td>13-39</td>\n",
       "      <td>3-7</td>\n",
       "      <td>3-7</td>\n",
       "      <td>6-4</td>\n",
       "      <td>...</td>\n",
       "      <td>6-24</td>\n",
       "      <td>5-4</td>\n",
       "      <td>9-39</td>\n",
       "      <td>0-1</td>\n",
       "      <td>3-14</td>\n",
       "      <td>7-9</td>\n",
       "      <td>6-6</td>\n",
       "      <td>5-7</td>\n",
       "      <td>2-15</td>\n",
       "      <td>2-5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>28</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>23-59</td>\n",
       "      <td>19-22</td>\n",
       "      <td>4-37</td>\n",
       "      <td>17-35</td>\n",
       "      <td>6-24</td>\n",
       "      <td>7-11</td>\n",
       "      <td>6-12</td>\n",
       "      <td>4-12</td>\n",
       "      <td>...</td>\n",
       "      <td>7-21</td>\n",
       "      <td>5-7</td>\n",
       "      <td>10-30</td>\n",
       "      <td>0-2</td>\n",
       "      <td>6-8</td>\n",
       "      <td>4-11</td>\n",
       "      <td>3-14</td>\n",
       "      <td>5-7</td>\n",
       "      <td>3-11</td>\n",
       "      <td>2-6</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>29</td>\n",
       "      <td>Philadelphia 76ers</td>\n",
       "      <td>19-63</td>\n",
       "      <td>10-31</td>\n",
       "      <td>9-32</td>\n",
       "      <td>14-38</td>\n",
       "      <td>5-25</td>\n",
       "      <td>5-11</td>\n",
       "      <td>4-14</td>\n",
       "      <td>5-13</td>\n",
       "      <td>...</td>\n",
       "      <td>4-24</td>\n",
       "      <td>6-3</td>\n",
       "      <td>4-46</td>\n",
       "      <td>1-0</td>\n",
       "      <td>5-11</td>\n",
       "      <td>3-10</td>\n",
       "      <td>6-11</td>\n",
       "      <td>0-11</td>\n",
       "      <td>1-15</td>\n",
       "      <td>3-5</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>30</td>\n",
       "      <td>Milwaukee Bucks</td>\n",
       "      <td>15-67</td>\n",
       "      <td>10-31</td>\n",
       "      <td>5-36</td>\n",
       "      <td>12-40</td>\n",
       "      <td>3-27</td>\n",
       "      <td>5-13</td>\n",
       "      <td>4-12</td>\n",
       "      <td>3-15</td>\n",
       "      <td>...</td>\n",
       "      <td>6-24</td>\n",
       "      <td>4-5</td>\n",
       "      <td>4-33</td>\n",
       "      <td>0-1</td>\n",
       "      <td>3-12</td>\n",
       "      <td>4-11</td>\n",
       "      <td>1-14</td>\n",
       "      <td>3-8</td>\n",
       "      <td>3-14</td>\n",
       "      <td>1-7</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>30 rows × 24 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "    Rk                    Team Overall   Home   Road      E      W     A  \\\n",
       "0    1       San Antonio Spurs   62-20   32-9  30-11   24-6  38-14   8-2   \n",
       "1    2   Oklahoma City Thunder   59-23   34-7  25-16   23-7  36-16   8-2   \n",
       "2    3    Los Angeles Clippers   57-25   34-7  23-18   21-9  36-16   9-1   \n",
       "3    4          Indiana Pacers   56-26   35-6  21-20  38-14  18-12  15-3   \n",
       "4    5         Houston Rockets   54-28   33-8  21-20   23-7  31-21   7-3   \n",
       "5    6              Miami Heat   54-28   32-9  22-19  34-18  20-10   9-9   \n",
       "6    7  Portland Trail Blazers   54-28  31-10  23-18   23-7  31-21   9-1   \n",
       "7    8   Golden State Warriors   51-31  27-14  24-17  20-10  31-21   7-3   \n",
       "8    9       Memphis Grizzlies   50-32  27-14  23-18   21-9  29-23   6-4   \n",
       "9   10        Dallas Mavericks   49-33  26-15  23-18  20-10  29-23   5-5   \n",
       "10  11           Chicago Bulls   48-34  27-14  21-20  35-17  13-17  12-6   \n",
       "11  12         Toronto Raptors   48-34  26-15  22-19  32-20  16-14  11-5   \n",
       "12  13            Phoenix Suns   48-34  26-15  22-19  20-10  28-24   7-3   \n",
       "13  14           Brooklyn Nets   44-38  28-13  16-25  26-26  18-12   9-7   \n",
       "14  15      Washington Wizards   44-38  22-19  22-19  33-19  11-19  13-5   \n",
       "15  16       Charlotte Bobcats   43-39  25-16  18-23  30-22  13-17  11-7   \n",
       "16  17  Minnesota Timberwolves   40-42  24-17  16-25  17-13  23-29   5-5   \n",
       "17  18           Atlanta Hawks   38-44  24-17  14-27  28-24  10-20  10-8   \n",
       "18  19         New York Knicks   37-45  19-22  18-23  26-26  11-19  10-6   \n",
       "19  20          Denver Nuggets   36-46  22-19  14-27  16-14  20-32   5-5   \n",
       "20  21    New Orleans Pelicans   34-48  22-19  12-29  19-11  15-37   6-4   \n",
       "21  22     Cleveland Cavaliers   33-49  19-22  14-27  21-31  12-18  8-10   \n",
       "22  23         Detroit Pistons   29-53  17-24  12-29  23-29   6-24  11-7   \n",
       "23  24        Sacramento Kings   28-54  17-24  11-30  13-17  15-37   5-5   \n",
       "24  25      Los Angeles Lakers   27-55  14-27  13-28  12-18  15-37   6-4   \n",
       "25  26          Boston Celtics   25-57  16-25   9-32  21-31   4-26  5-11   \n",
       "26  27               Utah Jazz   25-57  16-25   9-32  12-18  13-39   3-7   \n",
       "27  28           Orlando Magic   23-59  19-22   4-37  17-35   6-24  7-11   \n",
       "28  29      Philadelphia 76ers   19-63  10-31   9-32  14-38   5-25  5-11   \n",
       "29  30         Milwaukee Bucks   15-67  10-31   5-36  12-40   3-27  5-13   \n",
       "\n",
       "       C    SE  ...   Post     ≤3    ≥10  Oct   Nov   Dec   Jan   Feb   Mar  \\\n",
       "0    7-3   9-1  ...   24-5    7-0  35-12  1-0  13-3  11-4   8-6   9-3  16-0   \n",
       "1    8-2   7-3  ...  16-11   10-4   33-8  1-0  10-3  14-3  13-4   6-5  10-4   \n",
       "2    7-3   5-5  ...   20-7    4-3   30-9  1-1  11-4   9-7  12-4   7-4  13-2   \n",
       "3   12-4  11-7  ...  16-14    4-3  29-17  2-0  13-1  10-4  10-5   9-3  8-10   \n",
       "4    8-2   8-2  ...  18-11    7-6   30-9  1-0  12-5   8-8  10-4   8-2  10-4   \n",
       "5   13-5  12-4  ...  17-14    8-7  29-10  1-1  12-2  11-4   8-6   9-1  10-8   \n",
       "6    9-1   5-5  ...  18-11    8-9  24-13  0-1  13-2  12-4   8-6   7-5   8-9   \n",
       "7    7-3   6-4  ...   20-9   11-8   28-9  1-1   8-7  11-5   9-6   7-4   9-5   \n",
       "8    7-3   8-2  ...   21-9   10-4  18-13  0-1   8-7   5-9  12-3   7-5  12-5   \n",
       "9    9-1   6-4  ...  17-11  11-11  23-10  1-0   9-8   8-5   9-8   9-3   8-6   \n",
       "10  11-5  12-6  ...   21-9    6-5  27-19  1-1   6-7  5-10  11-4   9-4  10-6   \n",
       "11  13-5  8-10  ...  20-10    3-8   24-6  1-0   5-9   8-6  11-6   7-5  10-6   \n",
       "12   6-4   7-3  ...  18-13    2-9   19-7  1-0   8-8  10-3   9-7   6-6  10-6   \n",
       "13  7-11  10-8  ...  20-11    8-3  17-20  0-1  5-11   5-9  10-3   7-5  12-4   \n",
       "14  10-8  10-6  ...  19-11    5-9  19-14  0-1   8-8   6-5   8-9   8-5   8-8   \n",
       "15  13-5  6-10  ...   20-9    8-8  18-19  0-1   8-8   6-9   7-9   6-4   9-7   \n",
       "16   7-3   5-5  ...  15-14    5-6  26-18  1-0   8-9   6-7   8-7   5-6   8-8   \n",
       "17  10-8   8-8  ...  13-18    8-9  15-16  0-1   9-8   9-5   6-7  2-10  6-10   \n",
       "18   9-9  7-11  ...  17-13    3-8  19-22  1-1  2-11   6-9  10-6  2-11  11-5   \n",
       "19   5-5   6-4  ...  12-19    8-5  16-24  0-1   9-5  5-10   8-7   3-9  7-10   \n",
       "20   8-2   5-5  ...  11-19    7-8  10-21  0-1   7-7   7-7  5-11   4-9   9-7   \n",
       "21   7-9  6-12  ...  13-16    7-5  13-27  1-0  4-12   5-9   6-9   8-6   6-9   \n",
       "22  6-10  6-12  ...   7-23    2-6  16-25  1-0  5-10   8-9   4-8   5-8  4-12   \n",
       "23   4-6   4-6  ...  10-19    4-9  12-26  1-0  3-10  6-10  5-11   5-7  6-10   \n",
       "24   3-7   3-7  ...   9-20    5-5  12-33  1-1   8-7  4-11  3-12   4-8   5-9   \n",
       "25  6-12  10-8  ...   6-22    5-9   9-23  0-1  7-11   6-6  2-15   5-6  3-12   \n",
       "26   3-7   6-4  ...   6-24    5-4   9-39  0-1  3-14   7-9   6-6   5-7  2-15   \n",
       "27  6-12  4-12  ...   7-21    5-7  10-30  0-2   6-8  4-11  3-14   5-7  3-11   \n",
       "28  4-14  5-13  ...   4-24    6-3   4-46  1-0  5-11  3-10  6-11  0-11  1-15   \n",
       "29  4-12  3-15  ...   6-24    4-5   4-33  0-1  3-12  4-11  1-14   3-8  3-14   \n",
       "\n",
       "    Apr  \n",
       "0   4-4  \n",
       "1   5-4  \n",
       "2   4-3  \n",
       "3   4-3  \n",
       "4   5-5  \n",
       "5   3-6  \n",
       "6   6-1  \n",
       "7   6-3  \n",
       "8   6-2  \n",
       "9   5-3  \n",
       "10  6-2  \n",
       "11  6-2  \n",
       "12  4-4  \n",
       "13  5-5  \n",
       "14  6-2  \n",
       "15  7-1  \n",
       "16  4-5  \n",
       "17  6-3  \n",
       "18  5-2  \n",
       "19  4-4  \n",
       "20  2-6  \n",
       "21  3-4  \n",
       "22  2-6  \n",
       "23  2-6  \n",
       "24  2-7  \n",
       "25  2-6  \n",
       "26  2-5  \n",
       "27  2-6  \n",
       "28  3-5  \n",
       "29  1-7  \n",
       "\n",
       "[30 rows x 24 columns]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's try see which team is better on the ladder. Using the previous year's ladder\n",
    "ladder = pd.read_csv(NBA_EXTENDED_STANDINGS, skiprows=[0])\n",
    "ladder"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "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>Date</th>\n",
       "      <th>Start (ET)</th>\n",
       "      <th>Visitor Team</th>\n",
       "      <th>VisitorPts</th>\n",
       "      <th>Home Team</th>\n",
       "      <th>HomePts</th>\n",
       "      <th>Score Type</th>\n",
       "      <th>OT?</th>\n",
       "      <th>Attend.</th>\n",
       "      <th>LOG</th>\n",
       "      <th>Arena</th>\n",
       "      <th>Notes</th>\n",
       "      <th>HomeWin</th>\n",
       "      <th>HomeLastWin</th>\n",
       "      <th>VisitorLastWin</th>\n",
       "      <th>HomeWinStreak</th>\n",
       "      <th>VisitorWinStreak</th>\n",
       "      <th>HomeTeamRanksHigher</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>87</td>\n",
       "      <td>Indiana Pacers</td>\n",
       "      <td>97</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18165</td>\n",
       "      <td>2:17</td>\n",
       "      <td>Bankers Life Fieldhouse</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>10:30p</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>103</td>\n",
       "      <td>Los Angeles Lakers</td>\n",
       "      <td>116</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18997</td>\n",
       "      <td>2:27</td>\n",
       "      <td>STAPLES Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>8:00p</td>\n",
       "      <td>Chicago Bulls</td>\n",
       "      <td>95</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>107</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19964</td>\n",
       "      <td>2:32</td>\n",
       "      <td>AmericanAirlines Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>94</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>98</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20562</td>\n",
       "      <td>2:23</td>\n",
       "      <td>Quicken Loans Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>8:30p</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>109</td>\n",
       "      <td>Dallas Mavericks</td>\n",
       "      <td>118</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19834</td>\n",
       "      <td>2:14</td>\n",
       "      <td>American Airlines Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date Start (ET)          Visitor Team  VisitorPts  \\\n",
       "0 2013-10-29      7:00p         Orlando Magic          87   \n",
       "1 2013-10-29     10:30p  Los Angeles Clippers         103   \n",
       "2 2013-10-29      8:00p         Chicago Bulls          95   \n",
       "3 2013-10-30      7:00p         Brooklyn Nets          94   \n",
       "4 2013-10-30      8:30p         Atlanta Hawks         109   \n",
       "\n",
       "             Home Team  HomePts Score Type  OT?  Attend.   LOG  \\\n",
       "0       Indiana Pacers       97  Box Score  NaN    18165  2:17   \n",
       "1   Los Angeles Lakers      116  Box Score  NaN    18997  2:27   \n",
       "2           Miami Heat      107  Box Score  NaN    19964  2:32   \n",
       "3  Cleveland Cavaliers       98  Box Score  NaN    20562  2:23   \n",
       "4     Dallas Mavericks      118  Box Score  NaN    19834  2:14   \n",
       "\n",
       "                      Arena Notes  HomeWin HomeLastWin VisitorLastWin  \\\n",
       "0   Bankers Life Fieldhouse   NaN     True           0              0   \n",
       "1            STAPLES Center   NaN     True           0              0   \n",
       "2    AmericanAirlines Arena   NaN     True           0              0   \n",
       "3       Quicken Loans Arena   NaN     True           0              0   \n",
       "4  American Airlines Center   NaN     True           0              0   \n",
       "\n",
       "   HomeWinStreak  VisitorWinStreak  HomeTeamRanksHigher  \n",
       "0              0                 0                    0  \n",
       "1              0                 0                    1  \n",
       "2              0                 0                    0  \n",
       "3              0                 0                    1  \n",
       "4              0                 0                    0  "
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "\n",
    "# We can create a new feature -- HomeTeamRanksHigher\\\n",
    "results[\"HomeTeamRanksHigher\"] = 0\n",
    "for index, row in results.iterrows():\n",
    "    home_team = row[\"Home Team\"]\n",
    "    visitor_team = row[\"Visitor Team\"]\n",
    "    # if home_team == \"New Orleans Pelicans\":\n",
    "    #     home_team = \"New Orleans Hornets\"\n",
    "    # elif visitor_team == \"New Orleans Pelicans\":\n",
    "    #     visitor_team = \"New Orleans Hornets\"\n",
    "    home_rank = ladder[ladder[\"Team\"] == home_team][\"Rk\"].values[0]\n",
    "    visitor_rank = ladder[ladder[\"Team\"] == visitor_team][\"Rk\"].values[0]\n",
    "    row[\"HomeTeamRanksHigher\"] = int(home_rank > visitor_rank)\n",
    "    results.iloc[index] = row\n",
    "results.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using whether the home team is ranked higher\n",
      "Accuracy: 67.6%\n"
     ]
    }
   ],
   "source": [
    "X_homehigher =  results[[\"HomeLastWin\", \"VisitorLastWin\", \"HomeTeamRanksHigher\"]].values\n",
    "\n",
    "clf = DecisionTreeClassifier(random_state=14)\n",
    "scores = cross_val_score(clf, X_homehigher, y_true, scoring='accuracy')\n",
    "\n",
    "print(\"Using whether the home team is ranked higher\")\n",
    "print(f\"Accuracy: {np.mean(scores) * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This now scores 67.6 percent—even better than our previous result. Can we do better?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Extra: Model Training Using GridSearch"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 67.6%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.model_selection import GridSearchCV\n",
    "\n",
    "parameter_space = {\n",
    "                   \"max_depth\": [1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20],\n",
    "                   }\n",
    "clf = DecisionTreeClassifier(random_state=14)\n",
    "grid = GridSearchCV(clf, parameter_space)\n",
    "grid.fit(X_homehigher, y_true)\n",
    "print(f\"Accuracy: {grid.best_score_ * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Next, let's test which of the two teams won their last match. While rankings can give some hints on who won (the higher ranked team is more likely to win), sometimes teams play better against other teams. There are many reasons for this – for example,some teams may have strategies that work against other teams really well. Following our previous pattern, we create a dictionary to store the winner of the past game and create a new feature in our data frame. The code is as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "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>Date</th>\n",
       "      <th>Start (ET)</th>\n",
       "      <th>Visitor Team</th>\n",
       "      <th>VisitorPts</th>\n",
       "      <th>Home Team</th>\n",
       "      <th>HomePts</th>\n",
       "      <th>Score Type</th>\n",
       "      <th>OT?</th>\n",
       "      <th>Attend.</th>\n",
       "      <th>LOG</th>\n",
       "      <th>Arena</th>\n",
       "      <th>Notes</th>\n",
       "      <th>HomeWin</th>\n",
       "      <th>HomeLastWin</th>\n",
       "      <th>VisitorLastWin</th>\n",
       "      <th>HomeWinStreak</th>\n",
       "      <th>VisitorWinStreak</th>\n",
       "      <th>HomeTeamRanksHigher</th>\n",
       "      <th>HomeTeamWonLast</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Orlando Magic</td>\n",
       "      <td>87</td>\n",
       "      <td>Indiana Pacers</td>\n",
       "      <td>97</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18165</td>\n",
       "      <td>2:17</td>\n",
       "      <td>Bankers Life Fieldhouse</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>10:30p</td>\n",
       "      <td>Los Angeles Clippers</td>\n",
       "      <td>103</td>\n",
       "      <td>Los Angeles Lakers</td>\n",
       "      <td>116</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>18997</td>\n",
       "      <td>2:27</td>\n",
       "      <td>STAPLES Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2013-10-29</td>\n",
       "      <td>8:00p</td>\n",
       "      <td>Chicago Bulls</td>\n",
       "      <td>95</td>\n",
       "      <td>Miami Heat</td>\n",
       "      <td>107</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19964</td>\n",
       "      <td>2:32</td>\n",
       "      <td>AmericanAirlines Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>7:00p</td>\n",
       "      <td>Brooklyn Nets</td>\n",
       "      <td>94</td>\n",
       "      <td>Cleveland Cavaliers</td>\n",
       "      <td>98</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>20562</td>\n",
       "      <td>2:23</td>\n",
       "      <td>Quicken Loans Arena</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2013-10-30</td>\n",
       "      <td>8:30p</td>\n",
       "      <td>Atlanta Hawks</td>\n",
       "      <td>109</td>\n",
       "      <td>Dallas Mavericks</td>\n",
       "      <td>118</td>\n",
       "      <td>Box Score</td>\n",
       "      <td>NaN</td>\n",
       "      <td>19834</td>\n",
       "      <td>2:14</td>\n",
       "      <td>American Airlines Center</td>\n",
       "      <td>NaN</td>\n",
       "      <td>True</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        Date Start (ET)          Visitor Team  VisitorPts  \\\n",
       "0 2013-10-29      7:00p         Orlando Magic          87   \n",
       "1 2013-10-29     10:30p  Los Angeles Clippers         103   \n",
       "2 2013-10-29      8:00p         Chicago Bulls          95   \n",
       "3 2013-10-30      7:00p         Brooklyn Nets          94   \n",
       "4 2013-10-30      8:30p         Atlanta Hawks         109   \n",
       "\n",
       "             Home Team  HomePts Score Type  OT?  Attend.   LOG  \\\n",
       "0       Indiana Pacers       97  Box Score  NaN    18165  2:17   \n",
       "1   Los Angeles Lakers      116  Box Score  NaN    18997  2:27   \n",
       "2           Miami Heat      107  Box Score  NaN    19964  2:32   \n",
       "3  Cleveland Cavaliers       98  Box Score  NaN    20562  2:23   \n",
       "4     Dallas Mavericks      118  Box Score  NaN    19834  2:14   \n",
       "\n",
       "                      Arena Notes  HomeWin HomeLastWin VisitorLastWin  \\\n",
       "0   Bankers Life Fieldhouse   NaN     True           0              0   \n",
       "1            STAPLES Center   NaN     True           0              0   \n",
       "2    AmericanAirlines Arena   NaN     True           0              0   \n",
       "3       Quicken Loans Arena   NaN     True           0              0   \n",
       "4  American Airlines Center   NaN     True           0              0   \n",
       "\n",
       "   HomeWinStreak  VisitorWinStreak  HomeTeamRanksHigher  HomeTeamWonLast  \n",
       "0              0                 0                    0                0  \n",
       "1              0                 0                    1                0  \n",
       "2              0                 0                    0                0  \n",
       "3              0                 0                    1                0  \n",
       "4              0                 0                    0                0  "
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Who won the last match? We ignore home/visitor for this bit\n",
    "last_match_winner = defaultdict(int)\n",
    "results[\"HomeTeamWonLast\"] = 0\n",
    "\n",
    "for index, row in results.iterrows():\n",
    "    home_team = row[\"Home Team\"]\n",
    "    visitor_team = row[\"Visitor Team\"]\n",
    "    teams = tuple(sorted([home_team, visitor_team]))  # Sort for a consistent ordering\n",
    "    # Set in the row, who won the last encounter\n",
    "    row[\"HomeTeamWonLast\"] = 1 if last_match_winner[teams] == row[\"Home Team\"] else 0\n",
    "    results.iloc[index] = row\n",
    "    # Who won this one?\n",
    "    winner = row[\"Home Team\"] if row[\"HomeWin\"] else row[\"Visitor Team\"]\n",
    "    last_match_winner[teams] = winner\n",
    "results.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using whether the home team is ranked higher\n",
      "Accuracy: 67.6%\n"
     ]
    }
   ],
   "source": [
    "X_home_higher =  results[[\"HomeTeamRanksHigher\", \"HomeTeamWonLast\"]].values\n",
    "clf = DecisionTreeClassifier(random_state=14)\n",
    "scores = cross_val_score(clf, X_home_higher, y_true, scoring='accuracy')\n",
    "print(\"Using whether the home team is ranked higher\")\n",
    "print(f\"Accuracy: {np.mean(scores) * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "While decision trees are capable of learning from categorical features, the implementation in scikit-learn requires those features to be encoded first. We can use the LabelEncoder transformer to convert between the string-based team names into integers. The code is as follows:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 60.7%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.preprocessing import LabelEncoder, OneHotEncoder\n",
    "encoding = LabelEncoder()\n",
    "encoding.fit(results[\"Home Team\"].values)\n",
    "home_teams = encoding.transform(results[\"Home Team\"].values)\n",
    "visitor_teams = encoding.transform(results[\"Visitor Team\"].values)\n",
    "X_teams = np.vstack([home_teams, visitor_teams]).T\n",
    "\n",
    "onehot = OneHotEncoder()\n",
    "X_teams = onehot.fit_transform(X_teams).todense()\n",
    "\n",
    "clf = DecisionTreeClassifier(random_state=14)\n",
    "scores = cross_val_score(clf, np.asarray(X_teams), y_true, scoring='accuracy')\n",
    "print(f\"Accuracy: {np.mean(scores) * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This scores an accuracy of 60 percent. The score is better than the baseline, but not as good as before. It is possible that the larger number of features were not handled properly by the decision trees. For this reason, we will try changing the algorithm and see if that helps. Data mining can be an iterative process of trying new algorithms and features.\n",
    "\n",
    "## Random forests\n",
    "A single decision tree can learn quite complex functions. However, in many ways it will be prone to **overfitting—learning** (rules that work only for the training set). One of the ways that we can adjust for this is to limit the number of rules that it learns. For instance, we could limit the depth of the tree to just three layers. Such a tree will learn the best rules for splitting the dataset at a global level, but won't learn highly specific rules that separate the dataset into highly accurate groups.\n",
    "\n",
    "This trade-off results in trees that may have a good generalization, but overall slightly poorer performance.\n",
    "\n",
    "To compensate for this, we could create many decision trees and then ask each to predict the class value. We could take a majority vote and use that answer as our overall prediction. Random forests work on this principle. There are two problems with the aforementioned procedure. \n",
    "1. The first problem is that building decision trees is largely deterministic—using the same input will result in the same output each time. \n",
    "    - We only have one training dataset, which means our input (and therefore the output) will be the same if we try build multiple trees. We can address this by choosing a random subsample of our dataset, effectively creating new training sets. This process is called **bagging**.\n",
    "2. The second problem is that the features that are used for the first few decision nodes in our tree will be quite good. Even if we choose random subsamples of our training data, it is still quite possible that the decision trees built will be largely the same. \n",
    "    - To compensate for this, we also choose a random subset of the features to perform our data splits on.\n",
    "    \n",
    "Then, we have randomly built trees using randomly chosen samples, using (nearly) randomly chosen features. This is a Random Forest and, perhaps unintuitively, this algorithm is very effective on many datasets.\n",
    "\n",
    "## How do ensembles work?\n",
    "\n",
    "The randomness inherent in Random forests may make it seem like we are leaving the results of the algorithm up to chance. However, we apply the benefits of averaging to nearly randomly built decision trees, resulting in an algorithm that reduces the variance of the result.\n",
    "\n",
    "**Variance** is the error introduced by variations in the training dataset on the algorithm. Algorithms with a high variance (such as decision trees) can be greatly affected by variations to the training dataset. This results in models that have the problem of overfitting.\n",
    "\n",
    "**Note**:\n",
    "In contrast, **bias** is the error introduced by assumptions in the algorithm rather than anything to do with the dataset, that is, if we had an algorithm that presumed that all features would be normally distributed then our algorithm may have a high error if the features were not. Negative impacts from bias can be reduced by analyzing the data to see if the classifier's data model matches that of the actual data.\n",
    "\n",
    "In general, ensembles work on the assumption that errors in prediction are effectively random and that those errors are quite different from classifier to classifier. By averaging the results across many models, these random errors are canceled out—leaving the true prediction."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Parameters in Random forests\n",
    "\n",
    "The Random forest implementation in scikit-learn is called `RandomForestClassifier`, and it has a number of parameters. As Random forests use many instances of `DecisionTreeClassifier`, they share many of the same\n",
    "parameters such as the criterion (`Gini` Impurity or `Entropy/Information Gain`), `max_features`, and `min_samples_split`.\n",
    "\n",
    "Also, there are some new parameters that are used in the ensemble process:\n",
    "\n",
    "- *n_estimators*: This dictates how many decision trees should be built. A higher value will take longer to run, but will (probably) result in a higher accuracy.\n",
    "- *oob_score*: If true, the method is tested using samples that aren't in the random subsamples chosen for training the decision trees.\n",
    "- *n_jobs*: This specifies the number of cores to use when training the decision trees in parallel.\n",
    "\n",
    "The scikit-learn package uses a library called `Joblib` for in-built parallelization.This parameter dictates how many cores to use. By default, only a single core is used—if you have more cores, you can increase this, or set it to -1 to use all cores."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Applying Random forests\n",
    "\n",
    "Random forests in scikit-learn use the estimator interface, allowing us to use almost the exact same code as before to do cross fold validation:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using full team labels is ranked higher\n",
      "Accuracy: 61.4%\n"
     ]
    }
   ],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "clf = RandomForestClassifier(random_state=14)\n",
    "scores = cross_val_score(clf, np.asarray(X_teams), y_true, scoring='accuracy')\n",
    "print(\"Using full team labels is ranked higher\")\n",
    "print(f\"Accuracy: {np.mean(scores) * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This results in an immediate benefit of 61.4 percent, up by 1.3 points by just swapping the classifier.\n",
    "\n",
    "Random forests, using subsets of the features, should be able to learn more effectively with more features than normal decision trees. We can test this by throwing more features at the algorithm and seeing how it goes:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(1230, 62)\n"
     ]
    }
   ],
   "source": [
    "\n",
    "X_all = np.hstack([X_home_higher, X_teams])\n",
    "print(X_all.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Using whether the home team is ranked higher\n",
      "Accuracy: 62.3%\n"
     ]
    }
   ],
   "source": [
    "clf = RandomForestClassifier(random_state=14)\n",
    "scores = cross_val_score(clf, np.asarray(X_all), y_true, scoring='accuracy')\n",
    "print(\"Using whether the home team is ranked higher\")\n",
    "print(f\"Accuracy: {np.mean(scores) * 100:.1f}%\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This results in 62.3 percent —even better! We can also try some other parameters using the `GridSearchCV` Classifying with scikit-learn Estimators:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Accuracy: 66.0%\n",
      "RandomForestClassifier(criterion='entropy', max_features=10, min_samples_leaf=6,\n",
      "                       random_state=14)\n"
     ]
    }
   ],
   "source": [
    "#n_estimators=10, criterion='gini', max_depth=None, \n",
    "#min_samples_split=2, min_samples_leaf=1,\n",
    "#max_features='auto',\n",
    "#max_leaf_nodes=None, bootstrap=True,\n",
    "#oob_score=False, n_jobs=1,\n",
    "#random_state=None, verbose=0, min_density=None, compute_importances=None\n",
    "parameter_space = {\n",
    "                   \"max_features\": [2, 10, 'auto'],\n",
    "                   \"n_estimators\": [100,],\n",
    "                   \"criterion\": [\"gini\", \"entropy\"],\n",
    "                   \"min_samples_leaf\": [2, 4, 6],\n",
    "                   }\n",
    "clf = RandomForestClassifier(random_state=14)\n",
    "grid = GridSearchCV(clf, parameter_space)\n",
    "grid.fit(np.asarray(X_all), y_true)\n",
    "print(f\"Accuracy: {grid.best_score_ * 100:.1f}%\")\n",
    "print(grid.best_estimator_)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "This has a much better accuracy of 66 percent!\n",
    "\n",
    "## Engineering new features (a guide)\n",
    "\n",
    "In the previous few examples, we saw that changing the features can have quite a large impact on the performance of the algorithm. Through our small amount of testing, we had more than 10 percent variance just from the features. You can create features that come from a simple function in pandas by doing\n",
    "something like this:\n",
    "\n",
    "You can create features that come from a simple function in pandas by doing something like this:\n",
    "\n",
    "```python\n",
    "dataset[\"New Feature\"] = feature_creator()\n",
    "```\n",
    "\n",
    "The `feature_creator` function must return a list of the feature's value for each sample in the dataset. A common pattern is to use the dataset as a parameter:\n",
    "\n",
    "```python\n",
    "dataset[\"New Feature\"] = feature_creator(dataset)\n",
    "```\n",
    "\n",
    "You can create those features more directly by setting all the values to a single\n",
    "\"default\" value, like 0 in the next line:\n",
    "\n",
    "```python\n",
    "dataset[\"My New Feature\"] = 0\n",
    "```\n",
    "\n",
    "You can then iterate over the dataset, computing the features as you go. We used this format to create many of our features:\n",
    "\n",
    "```python\n",
    "for index, row in dataset.iterrows():\n",
    "home_team = row[\"Home Team\"]\n",
    "visitor_team = row[\"Visitor Team\"]\n",
    "# Some calculation here to alter row\n",
    "dataset.ix[index] = row\n",
    "```\n",
    "Keep in mind that this pattern isn't very efficient. If you are going to do this, try all of your features at once. A common \"best practice\" is to touch every sample as little as possible, preferably only once.\n",
    "\n",
    "Some example features that you could try and implement are as follows:\n",
    "- How many days has it been since each team's previous match? Teams may be tired if they play too many games in a short time frame.\n",
    "- How many games of the last five did each team win? This will give a more stable form of the `HomeLastWin` and `VisitorLastWin` features we extracted earlier (and can be extracted in a very similar way).\n",
    "- Do teams have a good record when visiting certain other teams? For instance, one team may play well in a particular stadium, even if they are the visitors.\n",
    "\n",
    "If you are facing trouble extracting features of these types, check the pandas documentation at http://pandas.pydata.org/pandas-docs/stable/ for help. Alternatively, you can try an online forum such as Stack Overflow for assistance.\n",
    "\n",
    "More extreme examples could use player data to estimate the strength of each team's sides to predict who won. These types of complex features are used every day by gamblers and sports betting agencies to try to turn a profit by predicting the outcome of sports matches."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  }
 ],
 "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": "3.7.0"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}