prosperloan-patch-3suwarna (3).ipynb

{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.7.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# Importing liabraries\nimport numpy as np\nimport pandas as pd\nimport matplotlib.pyplot as plt\nimport seaborn as sns\n%matplotlib inline","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:08.884445Z","iopub.execute_input":"2022-08-12T09:01:08.885018Z","iopub.status.idle":"2022-08-12T09:01:09.457489Z","shell.execute_reply.started":"2022-08-12T09:01:08.884888Z","shell.execute_reply":"2022-08-12T09:01:09.456533Z"},"trusted":true},"execution_count":1,"outputs":[]},{"cell_type":"code","source":"#import warnings\nimport warnings\nwarnings.filterwarnings('ignore')","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:09.459504Z","iopub.execute_input":"2022-08-12T09:01:09.459875Z","iopub.status.idle":"2022-08-12T09:01:09.465157Z","shell.execute_reply.started":"2022-08-12T09:01:09.459840Z","shell.execute_reply":"2022-08-12T09:01:09.463579Z"},"trusted":true},"execution_count":2,"outputs":[]},{"cell_type":"code","source":"df=pd.read_csv('../input/prosper-loan/prosperLoanData.csv')\ndf","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:09.466893Z","iopub.execute_input":"2022-08-12T09:01:09.467323Z","iopub.status.idle":"2022-08-12T09:01:11.863520Z","shell.execute_reply.started":"2022-08-12T09:01:09.467287Z","shell.execute_reply":"2022-08-12T09:01:11.862040Z"},"trusted":true},"execution_count":3,"outputs":[]},{"cell_type":"markdown","source":"# # Exploratory Data Analysis","metadata":{}},{"cell_type":"code","source":"##### What are the most number of borrowers Credit Grade?\n##### Check the univariate relationship of Credit Grade","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:11.866309Z","iopub.execute_input":"2022-08-12T09:01:11.867519Z","iopub.status.idle":"2022-08-12T09:01:11.873611Z","shell.execute_reply.started":"2022-08-12T09:01:11.867479Z","shell.execute_reply":"2022-08-12T09:01:11.872627Z"},"trusted":true},"execution_count":4,"outputs":[]},{"cell_type":"code","source":"#Borrower CreditGrade Distribution\nbase_color = sns.color_palette()[0]\ncat_order = df.CreditGrade.value_counts().index\nplt.figure(figsize=[8, 8])\nsns.countplot(data=df,y='CreditGrade',color=base_color, order=cat_order);\nplt.title('Borrower CreditGrade Distribution');\n\n\n\n##### CreditGrade 'C' has high Borrower CreditGrade.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:11.874900Z","iopub.execute_input":"2022-08-12T09:01:11.875374Z","iopub.status.idle":"2022-08-12T09:01:12.175068Z","shell.execute_reply.started":"2022-08-12T09:01:11.875336Z","shell.execute_reply":"2022-08-12T09:01:12.174144Z"},"trusted":true},"execution_count":5,"outputs":[]},{"cell_type":"code","source":"#####  Since there are so much low Credit Grade such as C and D , does it lead to a higher amount of deliquency?\n##### Check the univariate relationship of Loan Status\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.176426Z","iopub.execute_input":"2022-08-12T09:01:12.177498Z","iopub.status.idle":"2022-08-12T09:01:12.181984Z","shell.execute_reply.started":"2022-08-12T09:01:12.177451Z","shell.execute_reply":"2022-08-12T09:01:12.180950Z"},"trusted":true},"execution_count":6,"outputs":[]},{"cell_type":"code","source":"#Borrower CreditGrade Distribution\nbase_color = sns.color_palette()[0]\ncat_order = df.LoanStatus.value_counts().index\nplt.figure(figsize=[7, 5])\nsns.countplot(data=df,y='LoanStatus',color=base_color, order=cat_order);\nplt.title('LoanStatus Distribution');","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.183498Z","iopub.execute_input":"2022-08-12T09:01:12.183997Z","iopub.status.idle":"2022-08-12T09:01:12.468009Z","shell.execute_reply.started":"2022-08-12T09:01:12.183962Z","shell.execute_reply":"2022-08-12T09:01:12.467072Z"},"trusted":true},"execution_count":7,"outputs":[]},{"cell_type":"code","source":"#### What is the highest number of BorrowerRate?","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.469495Z","iopub.execute_input":"2022-08-12T09:01:12.471473Z","iopub.status.idle":"2022-08-12T09:01:12.475897Z","shell.execute_reply.started":"2022-08-12T09:01:12.471428Z","shell.execute_reply":"2022-08-12T09:01:12.474927Z"},"trusted":true},"execution_count":8,"outputs":[]},{"cell_type":"code","source":"plt.hist(df['BorrowerRate'])\n\n\n\n#####  The highest number of Borrower Rate is between 0.1 and 0.2","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.477803Z","iopub.execute_input":"2022-08-12T09:01:12.478247Z","iopub.status.idle":"2022-08-12T09:01:12.676915Z","shell.execute_reply.started":"2022-08-12T09:01:12.478199Z","shell.execute_reply":"2022-08-12T09:01:12.676061Z"},"trusted":true},"execution_count":9,"outputs":[]},{"cell_type":"code","source":"# Check the univariate relationship of Lender Yield on Loan\n#####  Since the highest number of Borrower Rate is between 0.1 and 0.2, does the highest number of Lender Yield is between 0.1 and 0.2?","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.681535Z","iopub.execute_input":"2022-08-12T09:01:12.682323Z","iopub.status.idle":"2022-08-12T09:01:12.686609Z","shell.execute_reply.started":"2022-08-12T09:01:12.682287Z","shell.execute_reply":"2022-08-12T09:01:12.685635Z"},"trusted":true},"execution_count":10,"outputs":[]},{"cell_type":"code","source":"plt.hist(df['LenderYield'])\n\n\n##### we see that Lender Yield is nomarlly distributed with most of the values between 0.1 and 0.2 \n#####   and we notice an increase at 0.3","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.687959Z","iopub.execute_input":"2022-08-12T09:01:12.688978Z","iopub.status.idle":"2022-08-12T09:01:12.885227Z","shell.execute_reply.started":"2022-08-12T09:01:12.688943Z","shell.execute_reply":"2022-08-12T09:01:12.884358Z"},"trusted":true},"execution_count":11,"outputs":[]},{"cell_type":"code","source":"##### Is the Credit Grade really accurate? Does higher Credit Grade leads to higher Monthly Loan Payment? As for Higher Credit Grade we mean from Grade AA to B\n##### # Check the Bivariate Relationship between CreditGarde and MonthlyLoan Payment\n#### Using Boxplot for numerical vs categorical variable.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.886751Z","iopub.execute_input":"2022-08-12T09:01:12.887093Z","iopub.status.idle":"2022-08-12T09:01:12.892398Z","shell.execute_reply.started":"2022-08-12T09:01:12.887058Z","shell.execute_reply":"2022-08-12T09:01:12.891262Z"},"trusted":true},"execution_count":12,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize = [12, 6])\nsns.boxplot(data=df,x='MonthlyLoanPayment',y='CreditGrade',color=base_color);\nplt.xlabel('MonthlyLoanPayment');\nplt.ylabel('CreditGrade');\nplt.title('MonthlyLoanPayment & CreditGrade');\n\n\n\n###### CreditGrade 'HR' & 'E' has low MonthlyLoanPayment as compared to others.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:12.894287Z","iopub.execute_input":"2022-08-12T09:01:12.895025Z","iopub.status.idle":"2022-08-12T09:01:13.320343Z","shell.execute_reply.started":"2022-08-12T09:01:12.894989Z","shell.execute_reply":"2022-08-12T09:01:13.319418Z"},"trusted":true},"execution_count":13,"outputs":[]},{"cell_type":"code","source":"#######  Here we look at the Completed Loan Status and Defaulted Rate to determine the accuracy of Credit Grade. \n###### Check the Bivariate Relatonship between CreditGrade and LoanStatus","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:13.321986Z","iopub.execute_input":"2022-08-12T09:01:13.322665Z","iopub.status.idle":"2022-08-12T09:01:13.327041Z","shell.execute_reply.started":"2022-08-12T09:01:13.322627Z","shell.execute_reply":"2022-08-12T09:01:13.326076Z"},"trusted":true},"execution_count":14,"outputs":[]},{"cell_type":"code","source":"##### One way of depicting the relationship between two categorical variables is through a heat map.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:13.328889Z","iopub.execute_input":"2022-08-12T09:01:13.329681Z","iopub.status.idle":"2022-08-12T09:01:13.343799Z","shell.execute_reply.started":"2022-08-12T09:01:13.329602Z","shell.execute_reply":"2022-08-12T09:01:13.342986Z"},"trusted":true},"execution_count":15,"outputs":[]},{"cell_type":"code","source":"\n##### The similar colors across the rows suggests that the \"LoanStatus\" group sizes are similarly sized and distributed similarly across levels of \"CreditGrade\". \n##### The 'Completed' LoanStatus has slightly higher counts of 'CreditGrade', 'C' observations.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:13.345181Z","iopub.execute_input":"2022-08-12T09:01:13.345806Z","iopub.status.idle":"2022-08-12T09:01:13.351116Z","shell.execute_reply.started":"2022-08-12T09:01:13.345756Z","shell.execute_reply":"2022-08-12T09:01:13.348793Z"},"trusted":true},"execution_count":16,"outputs":[]},{"cell_type":"code","source":"cat_counts = df.groupby(['CreditGrade', 'LoanStatus']).count()['LoanOriginationDate']\ncat_counts = cat_counts.reset_index()\ncat_counts = cat_counts.rename(columns={'LoanOriginationDate': 'count'}) \ncat_counts = cat_counts.pivot(index = 'CreditGrade', columns = 'LoanStatus', values = 'count')\ncat_counts.fillna(0, inplace=True)\n\nplt.figure(figsize = [10, 6])\nsns.heatmap(cat_counts, cmap = 'rocket_r', annot = True, fmt = \".0f\")\nplt.title('CreditGrade vs LoanStatus');","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:13.352780Z","iopub.execute_input":"2022-08-12T09:01:13.353894Z","iopub.status.idle":"2022-08-12T09:01:14.094915Z","shell.execute_reply.started":"2022-08-12T09:01:13.353861Z","shell.execute_reply":"2022-08-12T09:01:14.093936Z"},"trusted":true},"execution_count":17,"outputs":[]},{"cell_type":"code","source":"###### Now we know the Credit Grade is accurate and is a tool that is used by the organization in determining the person’s creditworthiness. Now we need to understand does the ProsperScore, the custom built risk assesment system is being used in determing borrower’s rate?\n######## # Check the Multivariate Relationship between BorrowerRate and BorrowerAPR.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:14.099495Z","iopub.execute_input":"2022-08-12T09:01:14.101864Z","iopub.status.idle":"2022-08-12T09:01:14.108265Z","shell.execute_reply.started":"2022-08-12T09:01:14.101828Z","shell.execute_reply":"2022-08-12T09:01:14.107295Z"},"trusted":true},"execution_count":18,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize = [10, 5])\nplt.scatter(data=df,x='BorrowerRate',y = 'BorrowerAPR',c='ProsperScore',cmap = 'viridis_r')\nplt.colorbar(label = 'ProsperScore');\nplt.xlabel('BorrowerRate')\nplt.ylabel('Borrower APR')\nplt.title('BorrowerRate and Borrower APR  vs ProsperRating');\n\n\n###### This chart confirms the relations between BorrowerAPR and BorrowerRate, they have strong positive relationship\n####### In addition, it shows that most of loans with lower BorrowerAPR & BorrowerRate have higher ProsperScore and visa-versa","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:14.111165Z","iopub.execute_input":"2022-08-12T09:01:14.112627Z","iopub.status.idle":"2022-08-12T09:01:15.781886Z","shell.execute_reply.started":"2022-08-12T09:01:14.112584Z","shell.execute_reply":"2022-08-12T09:01:15.780931Z"},"trusted":true},"execution_count":19,"outputs":[]},{"cell_type":"code","source":"##### ScorexChangeAtTimeOfListing and LoanStatus vs ProsperRating'\n\nplt.figure(figsize = [10, 5])\nplt.scatter(data=df,x='ScorexChangeAtTimeOfListing',y = 'LoanStatus',c='ProsperScore',cmap = 'viridis_r')\nplt.colorbar(label = 'ProsperScore');\nplt.xlabel('ScorexChangeAtTimeOfListing')\nplt.ylabel('LoanStatus')\nplt.title('ScorexChangeAtTimeOfListing and LoanStatus vs ProsperRating');\n\n\n#### Negative Relation between LoanStatus & ScorexChangeAtTimeOfListing and many LoanStatus has lower prosper score.\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:15.782891Z","iopub.execute_input":"2022-08-12T09:01:15.783257Z","iopub.status.idle":"2022-08-12T09:01:16.601410Z","shell.execute_reply.started":"2022-08-12T09:01:15.783206Z","shell.execute_reply":"2022-08-12T09:01:16.600474Z"},"trusted":true},"execution_count":20,"outputs":[]},{"cell_type":"code","source":"##### 'MonthlyLoanPayment and LoanStatus vs ProsperRating'\n\nplt.figure(figsize = [10, 5])\nplt.scatter(data=df,x='MonthlyLoanPayment',y = 'LoanStatus',c='ProsperScore',cmap = 'viridis_r')\nplt.colorbar(label = 'ProsperScore');\nplt.xlabel('MonthlyLoanPayment')\nplt.ylabel('LoanStatus')\nplt.title('MonthlyLoanPayment and LoanStatus vs ProsperRating');","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:16.602647Z","iopub.execute_input":"2022-08-12T09:01:16.603389Z","iopub.status.idle":"2022-08-12T09:01:18.648514Z","shell.execute_reply.started":"2022-08-12T09:01:16.603354Z","shell.execute_reply":"2022-08-12T09:01:18.647457Z"},"trusted":true},"execution_count":21,"outputs":[]},{"cell_type":"code","source":"#### DelinquenciesLast7Years\n\nplt.figure(figsize = [10, 5])\nplt.scatter(data=df,x='DelinquenciesLast7Years',y = 'LoanStatus',c='ProsperScore',cmap = 'viridis_r')\nplt.colorbar(label = 'ProsperScore');\nplt.xlabel('DelinquenciesLast7Years')\nplt.ylabel('LoanStatus')\nplt.title('DelinquenciesLast7Years and LoanStatus vs ProsperRating');\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:18.650303Z","iopub.execute_input":"2022-08-12T09:01:18.650667Z","iopub.status.idle":"2022-08-12T09:01:20.608803Z","shell.execute_reply.started":"2022-08-12T09:01:18.650633Z","shell.execute_reply":"2022-08-12T09:01:20.607727Z"},"trusted":true},"execution_count":22,"outputs":[]},{"cell_type":"code","source":"sns.heatmap(df.isnull(),yticklabels=False,cbar=False,cmap='viridis')","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:20.610933Z","iopub.execute_input":"2022-08-12T09:01:20.611320Z","iopub.status.idle":"2022-08-12T09:01:28.362242Z","shell.execute_reply.started":"2022-08-12T09:01:20.611285Z","shell.execute_reply":"2022-08-12T09:01:28.361271Z"},"trusted":true},"execution_count":23,"outputs":[]},{"cell_type":"code","source":"sns.set_style('darkgrid')\nsns.countplot(y='EmploymentStatus',data=df)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:28.363831Z","iopub.execute_input":"2022-08-12T09:01:28.364237Z","iopub.status.idle":"2022-08-12T09:01:28.793097Z","shell.execute_reply.started":"2022-08-12T09:01:28.364185Z","shell.execute_reply":"2022-08-12T09:01:28.792159Z"},"trusted":true},"execution_count":24,"outputs":[]},{"cell_type":"code","source":"sns.set_style('darkgrid')\nsns.countplot(y='LoanStatus',data=df)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:28.794449Z","iopub.execute_input":"2022-08-12T09:01:28.794888Z","iopub.status.idle":"2022-08-12T09:01:29.120575Z","shell.execute_reply.started":"2022-08-12T09:01:28.794851Z","shell.execute_reply":"2022-08-12T09:01:29.119532Z"},"trusted":true},"execution_count":25,"outputs":[]},{"cell_type":"code","source":"# Loan by term\nbase_color = sns.color_palette()[0]\nsns.countplot(data=df,x= 'Term',color=base_color);\nplt.title('Terms of loan (Months)')\nplt.xlabel('Term (Months)');\n\n\n##### loan term is either 1, 3 or 5 years so max 3 yrs loan is common.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:29.122080Z","iopub.execute_input":"2022-08-12T09:01:29.122397Z","iopub.status.idle":"2022-08-12T09:01:29.314412Z","shell.execute_reply.started":"2022-08-12T09:01:29.122370Z","shell.execute_reply":"2022-08-12T09:01:29.313509Z"},"trusted":true},"execution_count":26,"outputs":[]},{"cell_type":"code","source":"#Prosper Score Distribution\nbase_color = sns.color_palette()[0]\nsns.countplot(data=df,x= 'ProsperScore',color=base_color)\nplt.title('Prosper Score Distribution')\nplt.xlabel('Prosper Score');\n\n\n#### Prosper score of value 4 and 6 and 8 are the most common.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:29.315993Z","iopub.execute_input":"2022-08-12T09:01:29.316998Z","iopub.status.idle":"2022-08-12T09:01:29.569089Z","shell.execute_reply.started":"2022-08-12T09:01:29.316962Z","shell.execute_reply":"2022-08-12T09:01:29.568172Z"},"trusted":true},"execution_count":27,"outputs":[]},{"cell_type":"code","source":"#Borrower State Distrobution\ncat_order = df.BorrowerState.value_counts().index\nplt.figure(figsize=[8, 10])\nsns.countplot(data=df,y='BorrowerState',color=base_color, order=cat_order);\nplt.title('Borrower State Distribution');\n\n\n###### States California, Texas and New York have the 3 biggest numbers of loans originated in that period of time.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:29.576430Z","iopub.execute_input":"2022-08-12T09:01:29.577358Z","iopub.status.idle":"2022-08-12T09:01:30.218020Z","shell.execute_reply.started":"2022-08-12T09:01:29.577320Z","shell.execute_reply":"2022-08-12T09:01:30.217083Z"},"trusted":true},"execution_count":28,"outputs":[]},{"cell_type":"code","source":"#Borrower Occupation Distrobution\ncat_order = df.Occupation.value_counts().index\nplt.figure(figsize=[8, 13])\nsns.countplot(data=df,y='Occupation',color=base_color, order=cat_order);\nplt.title('Borrower Occupation Distribution');\n\n\n\n####### Most borrowers' occupations (excluding the others group) are: Professional, Computer Engineer and Executive\n ####### It is interesting to see that Students, Judges, Dentists and Biologists are the least occupations taking loans.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:30.219784Z","iopub.execute_input":"2022-08-12T09:01:30.220509Z","iopub.status.idle":"2022-08-12T09:01:31.262846Z","shell.execute_reply.started":"2022-08-12T09:01:30.220470Z","shell.execute_reply":"2022-08-12T09:01:31.259428Z"},"trusted":true},"execution_count":29,"outputs":[]},{"cell_type":"code","source":"# #Is Borrower-Homeowner Status Distrobution\nbase_color = sns.color_palette()[0]\nsns.countplot(data=df,x= 'IsBorrowerHomeowner',color=base_color);\nplt.title('Borrower-Homeowner Status Distrobution')\nplt.xlabel('IsBorrowerHomeowner');","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:31.264407Z","iopub.execute_input":"2022-08-12T09:01:31.264881Z","iopub.status.idle":"2022-08-12T09:01:31.455097Z","shell.execute_reply.started":"2022-08-12T09:01:31.264845Z","shell.execute_reply":"2022-08-12T09:01:31.454126Z"},"trusted":true},"execution_count":30,"outputs":[]},{"cell_type":"code","source":"# Changing Loan orgination date into date time format\ndf['LoanOriginationDate'] = pd.to_datetime(df['LoanOriginationDate'])","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:31.456752Z","iopub.execute_input":"2022-08-12T09:01:31.457093Z","iopub.status.idle":"2022-08-12T09:01:31.502455Z","shell.execute_reply.started":"2022-08-12T09:01:31.457059Z","shell.execute_reply":"2022-08-12T09:01:31.501554Z"},"trusted":true},"execution_count":31,"outputs":[]},{"cell_type":"code","source":"# Loan Origination by Year\nplt.figure(figsize=[8, 5])\nsns.countplot(data=df,x=df['LoanOriginationDate'].dt.year,color=base_color)\nplt.title('Loans Originated by Year')\nplt.xlabel('Loan Origination Year');","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:31.504222Z","iopub.execute_input":"2022-08-12T09:01:31.504778Z","iopub.status.idle":"2022-08-12T09:01:31.804297Z","shell.execute_reply.started":"2022-08-12T09:01:31.504742Z","shell.execute_reply":"2022-08-12T09:01:31.803235Z"},"trusted":true},"execution_count":32,"outputs":[]},{"cell_type":"code","source":"# Borrower APR vs Status of Loan and  Borrower APR vs Employment status\nplt.figure(figsize = [20, 5])\n\nplt.subplot(1, 2, 1)\nsns.boxplot(data=df,x='BorrowerAPR',y='LoanStatus',color=base_color);\nplt.xlabel('Borrower APR');\nplt.ylabel('Loan Status');\nplt.title('BorrowerAPR & LoanStatus');\n\nplt.subplot(1, 2, 2)\nsns.boxplot(data=df,x='BorrowerAPR',y='EmploymentStatus',color=base_color);\nplt.xlabel('Borrower APR');\nplt.ylabel('Employment Status');\nplt.title('BorrowerAPR & EmploymentStatus');\n\n\n##### On the left: Current and completed loans have lower rate than the past-due loans\n####### On the right: Employed people have lower rates than not employed","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:31.806291Z","iopub.execute_input":"2022-08-12T09:01:31.806954Z","iopub.status.idle":"2022-08-12T09:01:32.745008Z","shell.execute_reply.started":"2022-08-12T09:01:31.806916Z","shell.execute_reply":"2022-08-12T09:01:32.744145Z"},"trusted":true},"execution_count":33,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize = [12, 6])\nsns.boxplot(data=df,x='LoanOriginalAmount',y='EmploymentStatus',color=base_color);\nplt.xlabel('LoanOriginalAmount');\nplt.ylabel('Employment Status');\nplt.title('LoanOriginalAmount & EmploymentStatus');\n\n##### Employed people can get larger loan amounts compared to all other categories.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:32.746522Z","iopub.execute_input":"2022-08-12T09:01:32.748491Z","iopub.status.idle":"2022-08-12T09:01:33.114131Z","shell.execute_reply.started":"2022-08-12T09:01:32.748451Z","shell.execute_reply":"2022-08-12T09:01:33.113172Z"},"trusted":true},"execution_count":34,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# # ##### Data Cleaning (Feature Engg)","metadata":{}},{"cell_type":"code","source":"df.info()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:33.115734Z","iopub.execute_input":"2022-08-12T09:01:33.116078Z","iopub.status.idle":"2022-08-12T09:01:33.214337Z","shell.execute_reply.started":"2022-08-12T09:01:33.116043Z","shell.execute_reply":"2022-08-12T09:01:33.213271Z"},"trusted":true},"execution_count":35,"outputs":[]},{"cell_type":"code","source":"df.shape","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:33.215773Z","iopub.execute_input":"2022-08-12T09:01:33.216113Z","iopub.status.idle":"2022-08-12T09:01:33.224318Z","shell.execute_reply.started":"2022-08-12T09:01:33.216078Z","shell.execute_reply":"2022-08-12T09:01:33.222139Z"},"trusted":true},"execution_count":36,"outputs":[]},{"cell_type":"code","source":"# Duplicates data entry in loan data\ndf.duplicated().sum()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:33.225926Z","iopub.execute_input":"2022-08-12T09:01:33.226724Z","iopub.status.idle":"2022-08-12T09:01:33.644987Z","shell.execute_reply.started":"2022-08-12T09:01:33.226690Z","shell.execute_reply":"2022-08-12T09:01:33.643954Z"},"trusted":true},"execution_count":37,"outputs":[]},{"cell_type":"code","source":"# Changing Loan orgination date into date time format\ndf['LoanOriginationDate'] = pd.to_datetime(df['LoanOriginationDate'])","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:33.646560Z","iopub.execute_input":"2022-08-12T09:01:33.647205Z","iopub.status.idle":"2022-08-12T09:01:33.794479Z","shell.execute_reply.started":"2022-08-12T09:01:33.647161Z","shell.execute_reply":"2022-08-12T09:01:33.793374Z"},"trusted":true},"execution_count":38,"outputs":[]},{"cell_type":"code","source":"# Changing Loan orgination date into date time format\ndf['ListingCreationDate'] = pd.to_datetime(df['ListingCreationDate'])","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:33.795992Z","iopub.execute_input":"2022-08-12T09:01:33.796730Z","iopub.status.idle":"2022-08-12T09:01:33.848232Z","shell.execute_reply.started":"2022-08-12T09:01:33.796691Z","shell.execute_reply":"2022-08-12T09:01:33.847368Z"},"trusted":true},"execution_count":39,"outputs":[]},{"cell_type":"code","source":"#check missing value\ndf.isnull().sum()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:33.849641Z","iopub.execute_input":"2022-08-12T09:01:33.849988Z","iopub.status.idle":"2022-08-12T09:01:33.934114Z","shell.execute_reply.started":"2022-08-12T09:01:33.849954Z","shell.execute_reply":"2022-08-12T09:01:33.933080Z"},"trusted":true},"execution_count":40,"outputs":[]},{"cell_type":"code","source":"df.describe()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:33.935311Z","iopub.execute_input":"2022-08-12T09:01:33.935646Z","iopub.status.idle":"2022-08-12T09:01:34.362171Z","shell.execute_reply.started":"2022-08-12T09:01:33.935612Z","shell.execute_reply":"2022-08-12T09:01:34.361236Z"},"trusted":true},"execution_count":41,"outputs":[]},{"cell_type":"code","source":"# Checking Outlier\n# There is an Outlier\nsns.distplot(df['BorrowerAPR']) ","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:34.363788Z","iopub.execute_input":"2022-08-12T09:01:34.364421Z","iopub.status.idle":"2022-08-12T09:01:35.113003Z","shell.execute_reply.started":"2022-08-12T09:01:34.364382Z","shell.execute_reply":"2022-08-12T09:01:35.112081Z"},"trusted":true},"execution_count":42,"outputs":[]},{"cell_type":"code","source":"sns.distplot(df['EmploymentStatusDuration']) ","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:35.114723Z","iopub.execute_input":"2022-08-12T09:01:35.115470Z","iopub.status.idle":"2022-08-12T09:01:35.867635Z","shell.execute_reply.started":"2022-08-12T09:01:35.115429Z","shell.execute_reply":"2022-08-12T09:01:35.866158Z"},"trusted":true},"execution_count":43,"outputs":[]},{"cell_type":"code","source":"sns.distplot(df['ProsperPrincipalBorrowed']) ","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:35.870022Z","iopub.execute_input":"2022-08-12T09:01:35.871028Z","iopub.status.idle":"2022-08-12T09:01:36.344169Z","shell.execute_reply.started":"2022-08-12T09:01:35.870989Z","shell.execute_reply":"2022-08-12T09:01:36.343151Z"},"trusted":true},"execution_count":44,"outputs":[]},{"cell_type":"code","source":"sns.distplot(df['ProsperScore']) ","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:36.345713Z","iopub.execute_input":"2022-08-12T09:01:36.346100Z","iopub.status.idle":"2022-08-12T09:01:36.979444Z","shell.execute_reply.started":"2022-08-12T09:01:36.346065Z","shell.execute_reply":"2022-08-12T09:01:36.978468Z"},"trusted":true},"execution_count":45,"outputs":[]},{"cell_type":"code","source":"sns.distplot(df['EstimatedLoss'])","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:36.980773Z","iopub.execute_input":"2022-08-12T09:01:36.981153Z","iopub.status.idle":"2022-08-12T09:01:37.666261Z","shell.execute_reply.started":"2022-08-12T09:01:36.981114Z","shell.execute_reply":"2022-08-12T09:01:37.665132Z"},"trusted":true},"execution_count":46,"outputs":[]},{"cell_type":"code","source":"sns.boxplot(y='TotalProsperLoans',data=df)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:37.668370Z","iopub.execute_input":"2022-08-12T09:01:37.668758Z","iopub.status.idle":"2022-08-12T09:01:37.850071Z","shell.execute_reply.started":"2022-08-12T09:01:37.668721Z","shell.execute_reply":"2022-08-12T09:01:37.849135Z"},"trusted":true},"execution_count":47,"outputs":[]},{"cell_type":"code","source":"#categorical data (There is no outlier detection for categorical data)\ndf['CreditGrade'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:37.851584Z","iopub.execute_input":"2022-08-12T09:01:37.851937Z","iopub.status.idle":"2022-08-12T09:01:37.862967Z","shell.execute_reply.started":"2022-08-12T09:01:37.851903Z","shell.execute_reply":"2022-08-12T09:01:37.862049Z"},"trusted":true},"execution_count":48,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize = (13,4))\nsns.countplot(df.CreditGrade.dropna(), order = df.CreditGrade.value_counts().index);","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:37.864501Z","iopub.execute_input":"2022-08-12T09:01:37.865140Z","iopub.status.idle":"2022-08-12T09:01:38.110748Z","shell.execute_reply.started":"2022-08-12T09:01:37.865105Z","shell.execute_reply":"2022-08-12T09:01:38.109878Z"},"trusted":true},"execution_count":49,"outputs":[]},{"cell_type":"code","source":"df['LoanStatus'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.112274Z","iopub.execute_input":"2022-08-12T09:01:38.112632Z","iopub.status.idle":"2022-08-12T09:01:38.129026Z","shell.execute_reply.started":"2022-08-12T09:01:38.112598Z","shell.execute_reply":"2022-08-12T09:01:38.127397Z"},"trusted":true},"execution_count":50,"outputs":[]},{"cell_type":"code","source":"df['Occupation'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.130505Z","iopub.execute_input":"2022-08-12T09:01:38.130915Z","iopub.status.idle":"2022-08-12T09:01:38.149236Z","shell.execute_reply.started":"2022-08-12T09:01:38.130881Z","shell.execute_reply":"2022-08-12T09:01:38.148128Z"},"trusted":true},"execution_count":51,"outputs":[]},{"cell_type":"code","source":"df['EmploymentStatus'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.150783Z","iopub.execute_input":"2022-08-12T09:01:38.151852Z","iopub.status.idle":"2022-08-12T09:01:38.168173Z","shell.execute_reply.started":"2022-08-12T09:01:38.151811Z","shell.execute_reply":"2022-08-12T09:01:38.167130Z"},"trusted":true},"execution_count":52,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize = (13,4))\nsns.countplot(df.EmploymentStatus.dropna(), order = df.EmploymentStatus.value_counts().index);","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.169654Z","iopub.execute_input":"2022-08-12T09:01:38.170290Z","iopub.status.idle":"2022-08-12T09:01:38.452566Z","shell.execute_reply.started":"2022-08-12T09:01:38.170252Z","shell.execute_reply":"2022-08-12T09:01:38.451609Z"},"trusted":true},"execution_count":53,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize = (13,4))\nsns.countplot(df.IncomeVerifiable.dropna(), order = df.IncomeVerifiable.value_counts().index);","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.453931Z","iopub.execute_input":"2022-08-12T09:01:38.454933Z","iopub.status.idle":"2022-08-12T09:01:38.634102Z","shell.execute_reply.started":"2022-08-12T09:01:38.454888Z","shell.execute_reply":"2022-08-12T09:01:38.633118Z"},"trusted":true},"execution_count":54,"outputs":[]},{"cell_type":"code","source":"df['IncomeRange'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.635619Z","iopub.execute_input":"2022-08-12T09:01:38.635993Z","iopub.status.idle":"2022-08-12T09:01:38.651475Z","shell.execute_reply.started":"2022-08-12T09:01:38.635958Z","shell.execute_reply":"2022-08-12T09:01:38.650272Z"},"trusted":true},"execution_count":55,"outputs":[]},{"cell_type":"code","source":"plt.figure(figsize = (13,4))\nsns.countplot(df.IncomeRange.dropna(), order = df.IncomeRange.value_counts().index);","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.653423Z","iopub.execute_input":"2022-08-12T09:01:38.653992Z","iopub.status.idle":"2022-08-12T09:01:38.934320Z","shell.execute_reply.started":"2022-08-12T09:01:38.653958Z","shell.execute_reply":"2022-08-12T09:01:38.933406Z"},"trusted":true},"execution_count":56,"outputs":[]},{"cell_type":"code","source":"df_loan = df.copy()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.935752Z","iopub.execute_input":"2022-08-12T09:01:38.936093Z","iopub.status.idle":"2022-08-12T09:01:38.970980Z","shell.execute_reply.started":"2022-08-12T09:01:38.936058Z","shell.execute_reply":"2022-08-12T09:01:38.970042Z"},"trusted":true},"execution_count":57,"outputs":[]},{"cell_type":"code","source":"### Dropping column\ndf_loan.drop(['ListingCreationDate','LoanOriginationDate','GroupKey','CreditGrade','ProsperPrincipalBorrowed','ProsperPrincipalOutstanding','EstimatedEffectiveYield','EstimatedLoss','EstimatedReturn','ProsperRating (numeric)',\n        'TotalProsperLoans','TotalProsperPaymentsBilled','OnTimeProsperPayments','ProsperPaymentsLessThanOneMonthLate','ProsperPaymentsOneMonthPlusLate','ListingKey'], axis =1, inplace = True)\ndf_loan.head()\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:38.972357Z","iopub.execute_input":"2022-08-12T09:01:38.972813Z","iopub.status.idle":"2022-08-12T09:01:39.034472Z","shell.execute_reply.started":"2022-08-12T09:01:38.972772Z","shell.execute_reply":"2022-08-12T09:01:39.033269Z"},"trusted":true},"execution_count":58,"outputs":[]},{"cell_type":"code","source":"df_loan.info()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:39.035817Z","iopub.execute_input":"2022-08-12T09:01:39.037013Z","iopub.status.idle":"2022-08-12T09:01:39.115073Z","shell.execute_reply.started":"2022-08-12T09:01:39.036975Z","shell.execute_reply":"2022-08-12T09:01:39.113991Z"},"trusted":true},"execution_count":59,"outputs":[]},{"cell_type":"markdown","source":"### Handling missing values using imputation methods.\n#### Mean,Median,Mode","metadata":{}},{"cell_type":"code","source":"# All are categorical data so we use mode concept\ncols = ['ClosedDate','ProsperRating (Alpha)','BorrowerState','Occupation','EmploymentStatus','FirstRecordedCreditLine']\n\n\nfor i in cols:\n    df_loan[i].fillna(df_loan[i].mode()[0], inplace=True)\n    print(i)\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:39.116622Z","iopub.execute_input":"2022-08-12T09:01:39.116984Z","iopub.status.idle":"2022-08-12T09:01:39.198777Z","shell.execute_reply.started":"2022-08-12T09:01:39.116946Z","shell.execute_reply":"2022-08-12T09:01:39.197768Z"},"trusted":true},"execution_count":60,"outputs":[]},{"cell_type":"code","source":"### All are numerical columns so we use mean concept.\ncols =['BorrowerAPR','CreditScoreRangeLower','CreditScoreRangeUpper','ProsperScore','EmploymentStatusDuration','CurrentCreditLines','OpenCreditLines','TotalCreditLinespast7years','InquiriesLast6Months','TotalInquiries',\n      'CurrentDelinquencies','AmountDelinquent','DelinquenciesLast7Years','PublicRecordsLast10Years','PublicRecordsLast12Months','RevolvingCreditBalance','BankcardUtilization','AvailableBankcardCredit','TotalTrades',\n      'TradesNeverDelinquent (percentage)','TradesOpenedLast6Months','DebtToIncomeRatio','ScorexChangeAtTimeOfListing','LoanFirstDefaultedCycleNumber']\n\n\nfor i in cols:\n    df_loan[i].fillna(df_loan[i].mean(), inplace=True)\n    print(i)\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:39.200324Z","iopub.execute_input":"2022-08-12T09:01:39.200668Z","iopub.status.idle":"2022-08-12T09:01:39.233989Z","shell.execute_reply.started":"2022-08-12T09:01:39.200633Z","shell.execute_reply":"2022-08-12T09:01:39.232853Z"},"trusted":true},"execution_count":61,"outputs":[]},{"cell_type":"code","source":"# Correlation matrix for all numeric variables\nplt.figure(figsize = [10, 8])\nsns.heatmap(df_loan[cols].corr(), annot = True, fmt = '.3f',\n           cmap = 'vlag_r', center = 0)\nplt.title('Correlation Matrix') \nplt.margins(x = 0.5, y= 0.3)\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:39.235312Z","iopub.execute_input":"2022-08-12T09:01:39.235751Z","iopub.status.idle":"2022-08-12T09:01:42.013819Z","shell.execute_reply.started":"2022-08-12T09:01:39.235715Z","shell.execute_reply":"2022-08-12T09:01:42.011997Z"},"trusted":true},"execution_count":62,"outputs":[]},{"cell_type":"code","source":"df_loan.info()\n\n#### Now there is no null values present in dataset.","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:42.015249Z","iopub.execute_input":"2022-08-12T09:01:42.016339Z","iopub.status.idle":"2022-08-12T09:01:42.095660Z","shell.execute_reply.started":"2022-08-12T09:01:42.016301Z","shell.execute_reply":"2022-08-12T09:01:42.094524Z"},"trusted":true},"execution_count":63,"outputs":[]},{"cell_type":"markdown","source":"#### Encoding Concept \n#### For Categorical Columns","metadata":{}},{"cell_type":"code","source":"from sklearn.preprocessing import LabelEncoder\nle = LabelEncoder()\ncols = ['ClosedDate','ProsperRating (Alpha)','BorrowerState','Occupation','EmploymentStatus','FirstRecordedCreditLine',\n       'IsBorrowerHomeowner','CurrentlyInGroup','DateCreditPulled','IncomeRange','IncomeVerifiable','LoanKey','LoanOriginationQuarter',\n       'MemberKey']\n\n# Encode labels of multiple columns at once\n\ndf_loan[cols] = df_loan[cols].apply(LabelEncoder().fit_transform)\n\n# Print head\n\ndf_loan.head()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:42.096980Z","iopub.execute_input":"2022-08-12T09:01:42.097808Z","iopub.status.idle":"2022-08-12T09:01:43.406059Z","shell.execute_reply.started":"2022-08-12T09:01:42.097779Z","shell.execute_reply":"2022-08-12T09:01:43.404879Z"},"trusted":true},"execution_count":64,"outputs":[]},{"cell_type":"code","source":"df_loan.info()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.410552Z","iopub.execute_input":"2022-08-12T09:01:43.413256Z","iopub.status.idle":"2022-08-12T09:01:43.465313Z","shell.execute_reply.started":"2022-08-12T09:01:43.413202Z","shell.execute_reply":"2022-08-12T09:01:43.462419Z"},"trusted":true},"execution_count":65,"outputs":[]},{"cell_type":"code","source":"### Building a Model\nX = df_loan.drop(columns='LoanStatus', axis=1)\nY = df_loan['LoanStatus']","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.466875Z","iopub.execute_input":"2022-08-12T09:01:43.467495Z","iopub.status.idle":"2022-08-12T09:01:43.513311Z","shell.execute_reply.started":"2022-08-12T09:01:43.467456Z","shell.execute_reply":"2022-08-12T09:01:43.512255Z"},"trusted":true},"execution_count":66,"outputs":[]},{"cell_type":"code","source":"X","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.514870Z","iopub.execute_input":"2022-08-12T09:01:43.515276Z","iopub.status.idle":"2022-08-12T09:01:43.567529Z","shell.execute_reply.started":"2022-08-12T09:01:43.515238Z","shell.execute_reply":"2022-08-12T09:01:43.566523Z"},"trusted":true},"execution_count":67,"outputs":[]},{"cell_type":"code","source":"df_loan['LoanStatus']","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.569897Z","iopub.execute_input":"2022-08-12T09:01:43.570177Z","iopub.status.idle":"2022-08-12T09:01:43.577865Z","shell.execute_reply.started":"2022-08-12T09:01:43.570151Z","shell.execute_reply":"2022-08-12T09:01:43.576906Z"},"trusted":true},"execution_count":68,"outputs":[]},{"cell_type":"code","source":"#Encoding concept (bec of catogorical data)\ndf_loan1 = pd.get_dummies(df_loan['LoanStatus'])\nprint(df_loan1.head())\n\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.579438Z","iopub.execute_input":"2022-08-12T09:01:43.580043Z","iopub.status.idle":"2022-08-12T09:01:43.607852Z","shell.execute_reply.started":"2022-08-12T09:01:43.580007Z","shell.execute_reply":"2022-08-12T09:01:43.606945Z"},"trusted":true},"execution_count":69,"outputs":[]},{"cell_type":"code","source":"#####df_loan['LoanStatus'].value_counts()\ndf['LoanStatus'].value_counts()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.609284Z","iopub.execute_input":"2022-08-12T09:01:43.609658Z","iopub.status.idle":"2022-08-12T09:01:43.626821Z","shell.execute_reply.started":"2022-08-12T09:01:43.609619Z","shell.execute_reply":"2022-08-12T09:01:43.625798Z"},"trusted":true},"execution_count":70,"outputs":[]},{"cell_type":"code","source":"# display result\ndf_loan2 = pd.concat((df_loan1, df_loan), axis=1)\ndf_loan2 = df_loan2.drop([\"LoanStatus\"], axis=1)\ndf_loan2 = df_loan2.drop([\"Cancelled\",\"Chargedoff\",\"Current\",\"Defaulted\",\"FinalPaymentInProgress\",\"FinalPaymentInProgress\",\"Past Due (1-15 days)\",\"Past Due (16-30 days)\",\n                              \"Past Due (31-60 days)\",\"Past Due (61-90 days)\",\"Past Due (91-120 days)\",\"Past Due (91-120 days)\",\"Past Due (>120 days)\"], axis=1)\ndata= df_loan2.rename(columns={\"Completed\": \"LoanStatus\"})\nprint(data)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.628488Z","iopub.execute_input":"2022-08-12T09:01:43.628851Z","iopub.status.idle":"2022-08-12T09:01:43.763548Z","shell.execute_reply.started":"2022-08-12T09:01:43.628818Z","shell.execute_reply":"2022-08-12T09:01:43.762506Z"},"trusted":true},"execution_count":71,"outputs":[]},{"cell_type":"code","source":"data","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.764942Z","iopub.execute_input":"2022-08-12T09:01:43.765338Z","iopub.status.idle":"2022-08-12T09:01:43.809920Z","shell.execute_reply.started":"2022-08-12T09:01:43.765302Z","shell.execute_reply":"2022-08-12T09:01:43.808832Z"},"trusted":true},"execution_count":72,"outputs":[]},{"cell_type":"code","source":"### Building a Model\nX = data.drop(columns='LoanStatus', axis=1)\nY = data['LoanStatus']","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.811124Z","iopub.execute_input":"2022-08-12T09:01:43.811536Z","iopub.status.idle":"2022-08-12T09:01:43.836885Z","shell.execute_reply.started":"2022-08-12T09:01:43.811500Z","shell.execute_reply":"2022-08-12T09:01:43.835873Z"},"trusted":true},"execution_count":73,"outputs":[]},{"cell_type":"code","source":"X","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.838496Z","iopub.execute_input":"2022-08-12T09:01:43.839113Z","iopub.status.idle":"2022-08-12T09:01:43.894685Z","shell.execute_reply.started":"2022-08-12T09:01:43.839075Z","shell.execute_reply":"2022-08-12T09:01:43.893739Z"},"trusted":true},"execution_count":74,"outputs":[]},{"cell_type":"code","source":"Y","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.896070Z","iopub.execute_input":"2022-08-12T09:01:43.897180Z","iopub.status.idle":"2022-08-12T09:01:43.905277Z","shell.execute_reply.started":"2022-08-12T09:01:43.897140Z","shell.execute_reply":"2022-08-12T09:01:43.904158Z"},"trusted":true},"execution_count":75,"outputs":[]},{"cell_type":"code","source":"### Splitting the data into training data & Testing data\nfrom sklearn.model_selection import train_test_split\nX_train, X_test, Y_train, Y_test = train_test_split(X, Y, test_size=0.20, random_state=100)\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:43.906497Z","iopub.execute_input":"2022-08-12T09:01:43.907477Z","iopub.status.idle":"2022-08-12T09:01:44.017560Z","shell.execute_reply.started":"2022-08-12T09:01:43.907435Z","shell.execute_reply":"2022-08-12T09:01:44.016571Z"},"trusted":true},"execution_count":76,"outputs":[]},{"cell_type":"code","source":"print(X.shape, X_train.shape, X_test.shape)\n","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:44.018890Z","iopub.execute_input":"2022-08-12T09:01:44.019350Z","iopub.status.idle":"2022-08-12T09:01:44.028540Z","shell.execute_reply.started":"2022-08-12T09:01:44.019309Z","shell.execute_reply":"2022-08-12T09:01:44.025789Z"},"trusted":true},"execution_count":77,"outputs":[]},{"cell_type":"markdown","source":"# # ##### Last step is Scalling part in FE (Standarisation & Normalization)","metadata":{}},{"cell_type":"markdown","source":"# ### Feature Selection ","metadata":{}},{"cell_type":"code","source":"from sklearn.feature_selection import mutual_info_classif\n# determine the mutual information\nmutual_info = mutual_info_classif(X_train, Y_train)\nmutual_info","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:01:44.030187Z","iopub.execute_input":"2022-08-12T09:01:44.030543Z","iopub.status.idle":"2022-08-12T09:02:12.504188Z","shell.execute_reply.started":"2022-08-12T09:01:44.030501Z","shell.execute_reply":"2022-08-12T09:02:12.503156Z"},"trusted":true},"execution_count":78,"outputs":[]},{"cell_type":"code","source":"mutual_info = pd.Series(mutual_info)\nmutual_info.index = X_train.columns\nmutual_info.sort_values(ascending=False)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:12.505699Z","iopub.execute_input":"2022-08-12T09:02:12.506071Z","iopub.status.idle":"2022-08-12T09:02:12.515923Z","shell.execute_reply.started":"2022-08-12T09:02:12.506042Z","shell.execute_reply":"2022-08-12T09:02:12.514702Z"},"trusted":true},"execution_count":79,"outputs":[]},{"cell_type":"code","source":"#let's plot the ordered mutual_info values per feature\nmutual_info.sort_values(ascending=False).plot.bar(figsize=(20, 8))","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:12.517990Z","iopub.execute_input":"2022-08-12T09:02:12.518718Z","iopub.status.idle":"2022-08-12T09:02:14.714318Z","shell.execute_reply.started":"2022-08-12T09:02:12.518680Z","shell.execute_reply":"2022-08-12T09:02:14.713263Z"},"trusted":true},"execution_count":80,"outputs":[]},{"cell_type":"code","source":"from sklearn.feature_selection import SelectKBest","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:14.715863Z","iopub.execute_input":"2022-08-12T09:02:14.716299Z","iopub.status.idle":"2022-08-12T09:02:14.721980Z","shell.execute_reply.started":"2022-08-12T09:02:14.716266Z","shell.execute_reply":"2022-08-12T09:02:14.720888Z"},"trusted":true},"execution_count":81,"outputs":[]},{"cell_type":"code","source":"#Now we Will select the  top 5 important features\nsel_five_cols = SelectKBest(mutual_info_classif, k=10)\nsel_five_cols.fit(X_train, Y_train)\nX_train.columns[sel_five_cols.get_support()]","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:14.735148Z","iopub.execute_input":"2022-08-12T09:02:14.735432Z","iopub.status.idle":"2022-08-12T09:02:43.505582Z","shell.execute_reply.started":"2022-08-12T09:02:14.735406Z","shell.execute_reply":"2022-08-12T09:02:43.504292Z"},"trusted":true},"execution_count":82,"outputs":[]},{"cell_type":"code","source":"X_train_new = sel_five_cols.transform(X_train)\nX_test_new =  sel_five_cols.transform(X_test)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:43.507378Z","iopub.execute_input":"2022-08-12T09:02:43.508237Z","iopub.status.idle":"2022-08-12T09:02:43.544791Z","shell.execute_reply.started":"2022-08-12T09:02:43.508179Z","shell.execute_reply":"2022-08-12T09:02:43.543800Z"},"trusted":true},"execution_count":83,"outputs":[]},{"cell_type":"markdown","source":"# #### Standardize the data","metadata":{}},{"cell_type":"code","source":"from sklearn.preprocessing import StandardScaler","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:43.546323Z","iopub.execute_input":"2022-08-12T09:02:43.546933Z","iopub.status.idle":"2022-08-12T09:02:43.551836Z","shell.execute_reply.started":"2022-08-12T09:02:43.546895Z","shell.execute_reply":"2022-08-12T09:02:43.550816Z"},"trusted":true},"execution_count":84,"outputs":[]},{"cell_type":"code","source":"scaler = StandardScaler()\nX_train_std = scaler.fit_transform(X_train_new)\nX_test_std = scaler.transform(X_test_new)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:43.553836Z","iopub.execute_input":"2022-08-12T09:02:43.554546Z","iopub.status.idle":"2022-08-12T09:02:43.571771Z","shell.execute_reply.started":"2022-08-12T09:02:43.554511Z","shell.execute_reply":"2022-08-12T09:02:43.570890Z"},"trusted":true},"execution_count":85,"outputs":[]},{"cell_type":"markdown","source":"# ###### Model Creation","metadata":{}},{"cell_type":"code","source":"from sklearn.linear_model import LogisticRegression\nclf = LogisticRegression()\n# training the model\nclf.fit(X_train_std, Y_train)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:43.572866Z","iopub.execute_input":"2022-08-12T09:02:43.574435Z","iopub.status.idle":"2022-08-12T09:02:44.528628Z","shell.execute_reply.started":"2022-08-12T09:02:43.574400Z","shell.execute_reply":"2022-08-12T09:02:44.527380Z"},"trusted":true},"execution_count":86,"outputs":[]},{"cell_type":"code","source":"y_pred = clf.predict(X_test_std)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.534580Z","iopub.execute_input":"2022-08-12T09:02:44.535316Z","iopub.status.idle":"2022-08-12T09:02:44.543971Z","shell.execute_reply.started":"2022-08-12T09:02:44.535277Z","shell.execute_reply":"2022-08-12T09:02:44.542712Z"},"trusted":true},"execution_count":87,"outputs":[]},{"cell_type":"code","source":"from sklearn.metrics import accuracy_score\naccuracy = accuracy_score(Y_test, y_pred)\nprint(\"Accuracy: {:.2f}%\".format(accuracy * 100))","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.545505Z","iopub.execute_input":"2022-08-12T09:02:44.546082Z","iopub.status.idle":"2022-08-12T09:02:44.555544Z","shell.execute_reply.started":"2022-08-12T09:02:44.546046Z","shell.execute_reply":"2022-08-12T09:02:44.554570Z"},"trusted":true},"execution_count":88,"outputs":[]},{"cell_type":"code","source":"from sklearn.metrics import confusion_matrix\ncf = confusion_matrix(Y_test, y_pred)\nplt.figure()\nsns.heatmap(cf, annot=True)\nplt.xlabel('Prediction')\nplt.ylabel('Target')\nplt.title('Confusion Matrix')","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.556913Z","iopub.execute_input":"2022-08-12T09:02:44.557639Z","iopub.status.idle":"2022-08-12T09:02:44.838217Z","shell.execute_reply.started":"2022-08-12T09:02:44.557605Z","shell.execute_reply":"2022-08-12T09:02:44.837134Z"},"trusted":true},"execution_count":89,"outputs":[]},{"cell_type":"code","source":"# Evaluate the model\nfrom sklearn.metrics import classification_report, confusion_matrix","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.839913Z","iopub.execute_input":"2022-08-12T09:02:44.840327Z","iopub.status.idle":"2022-08-12T09:02:44.845265Z","shell.execute_reply.started":"2022-08-12T09:02:44.840291Z","shell.execute_reply":"2022-08-12T09:02:44.844135Z"},"trusted":true},"execution_count":90,"outputs":[]},{"cell_type":"code","source":"print(classification_report(Y_test,y_pred))","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.847005Z","iopub.execute_input":"2022-08-12T09:02:44.847403Z","iopub.status.idle":"2022-08-12T09:02:44.892216Z","shell.execute_reply.started":"2022-08-12T09:02:44.847366Z","shell.execute_reply":"2022-08-12T09:02:44.891163Z"},"trusted":true},"execution_count":91,"outputs":[]},{"cell_type":"code","source":"confusion_matrix(Y_test,y_pred)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.893722Z","iopub.execute_input":"2022-08-12T09:02:44.894418Z","iopub.status.idle":"2022-08-12T09:02:44.906351Z","shell.execute_reply.started":"2022-08-12T09:02:44.894378Z","shell.execute_reply":"2022-08-12T09:02:44.905278Z"},"trusted":true},"execution_count":92,"outputs":[]},{"cell_type":"code","source":"(14281+6794)/(14281+898+815+6794)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.907653Z","iopub.execute_input":"2022-08-12T09:02:44.908078Z","iopub.status.idle":"2022-08-12T09:02:44.915678Z","shell.execute_reply.started":"2022-08-12T09:02:44.908042Z","shell.execute_reply":"2022-08-12T09:02:44.914474Z"},"trusted":true},"execution_count":93,"outputs":[]},{"cell_type":"markdown","source":"# 92% ACCURACY (Logistics Regression)","metadata":{}},{"cell_type":"markdown","source":"# #### Decision Tree Algorithm (98% Accuracy)","metadata":{}},{"cell_type":"code","source":"from sklearn.tree import DecisionTreeClassifier\nmodel = DecisionTreeClassifier(criterion = \"entropy\")","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.916993Z","iopub.execute_input":"2022-08-12T09:02:44.918166Z","iopub.status.idle":"2022-08-12T09:02:44.945668Z","shell.execute_reply.started":"2022-08-12T09:02:44.918128Z","shell.execute_reply":"2022-08-12T09:02:44.944817Z"},"trusted":true},"execution_count":94,"outputs":[]},{"cell_type":"code","source":"model.fit(X_train_std, Y_train)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:44.946831Z","iopub.execute_input":"2022-08-12T09:02:44.947795Z","iopub.status.idle":"2022-08-12T09:02:45.647822Z","shell.execute_reply.started":"2022-08-12T09:02:44.947760Z","shell.execute_reply":"2022-08-12T09:02:45.646697Z"},"trusted":true},"execution_count":95,"outputs":[]},{"cell_type":"code","source":"y_pred = model.predict(X_test_std)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:45.649656Z","iopub.execute_input":"2022-08-12T09:02:45.650064Z","iopub.status.idle":"2022-08-12T09:02:45.658186Z","shell.execute_reply.started":"2022-08-12T09:02:45.650023Z","shell.execute_reply":"2022-08-12T09:02:45.657296Z"},"trusted":true},"execution_count":96,"outputs":[]},{"cell_type":"code","source":"from sklearn.metrics import accuracy_score\naccuracy = accuracy_score(Y_test, y_pred)\nprint(\"Accuracy: {:.2f}%\".format(accuracy * 100))","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:45.660003Z","iopub.execute_input":"2022-08-12T09:02:45.660425Z","iopub.status.idle":"2022-08-12T09:02:45.670641Z","shell.execute_reply.started":"2022-08-12T09:02:45.660395Z","shell.execute_reply":"2022-08-12T09:02:45.669500Z"},"trusted":true},"execution_count":97,"outputs":[]},{"cell_type":"code","source":"# Evaluate the model\nfrom sklearn.metrics import classification_report, confusion_matrix","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:45.672149Z","iopub.execute_input":"2022-08-12T09:02:45.672820Z","iopub.status.idle":"2022-08-12T09:02:45.678809Z","shell.execute_reply.started":"2022-08-12T09:02:45.672778Z","shell.execute_reply":"2022-08-12T09:02:45.677658Z"},"trusted":true},"execution_count":98,"outputs":[]},{"cell_type":"code","source":"print(classification_report(Y_test,y_pred))","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:02:45.680692Z","iopub.execute_input":"2022-08-12T09:02:45.681506Z","iopub.status.idle":"2022-08-12T09:02:45.728447Z","shell.execute_reply.started":"2022-08-12T09:02:45.681467Z","shell.execute_reply":"2022-08-12T09:02:45.727617Z"},"trusted":true},"execution_count":99,"outputs":[]},{"cell_type":"markdown","source":"# ##### Naives Bayes ","metadata":{}},{"cell_type":"code","source":"from sklearn.naive_bayes import GaussianNB\ngnb = GaussianNB()\ngnb.fit(X_train_std, Y_train)\n ","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:05:17.828596Z","iopub.execute_input":"2022-08-12T09:05:17.829072Z","iopub.status.idle":"2022-08-12T09:05:17.873766Z","shell.execute_reply.started":"2022-08-12T09:05:17.829030Z","shell.execute_reply":"2022-08-12T09:05:17.872744Z"},"trusted":true},"execution_count":101,"outputs":[]},{"cell_type":"code","source":"y_pred = gnb.predict(X_test_std)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:05:30.972754Z","iopub.execute_input":"2022-08-12T09:05:30.973118Z","iopub.status.idle":"2022-08-12T09:05:30.981710Z","shell.execute_reply.started":"2022-08-12T09:05:30.973085Z","shell.execute_reply":"2022-08-12T09:05:30.980611Z"},"trusted":true},"execution_count":103,"outputs":[]},{"cell_type":"code","source":"from sklearn import metrics\nprint(\"Gaussian Naive Bayes model accuracy(in %):\", metrics.accuracy_score(Y_test, y_pred)*100)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:05:36.838317Z","iopub.execute_input":"2022-08-12T09:05:36.838925Z","iopub.status.idle":"2022-08-12T09:05:36.848329Z","shell.execute_reply.started":"2022-08-12T09:05:36.838888Z","shell.execute_reply":"2022-08-12T09:05:36.846847Z"},"trusted":true},"execution_count":104,"outputs":[]},{"cell_type":"markdown","source":"# Save the Model For Deployment","metadata":{}},{"cell_type":"code","source":"model = GaussianNB()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:12:44.981184Z","iopub.execute_input":"2022-08-12T09:12:44.981795Z","iopub.status.idle":"2022-08-12T09:12:44.986507Z","shell.execute_reply.started":"2022-08-12T09:12:44.981760Z","shell.execute_reply":"2022-08-12T09:12:44.985193Z"},"trusted":true},"execution_count":105,"outputs":[]},{"cell_type":"code","source":"model.fit(X_train_std, Y_train)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:12:52.594083Z","iopub.execute_input":"2022-08-12T09:12:52.595081Z","iopub.status.idle":"2022-08-12T09:12:52.623646Z","shell.execute_reply.started":"2022-08-12T09:12:52.595034Z","shell.execute_reply":"2022-08-12T09:12:52.622579Z"},"trusted":true},"execution_count":106,"outputs":[]},{"cell_type":"code","source":"y_pred = model.predict(X_test_std)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:12:55.452617Z","iopub.execute_input":"2022-08-12T09:12:55.452980Z","iopub.status.idle":"2022-08-12T09:12:55.462114Z","shell.execute_reply.started":"2022-08-12T09:12:55.452947Z","shell.execute_reply":"2022-08-12T09:12:55.461060Z"},"trusted":true},"execution_count":107,"outputs":[]},{"cell_type":"code","source":"score = accuracy_score(Y_test, y_pred)\nprint(\"Accuracy: \", score)","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:14:04.630838Z","iopub.execute_input":"2022-08-12T09:14:04.631232Z","iopub.status.idle":"2022-08-12T09:14:04.639181Z","shell.execute_reply.started":"2022-08-12T09:14:04.631179Z","shell.execute_reply":"2022-08-12T09:14:04.638214Z"},"trusted":true},"execution_count":108,"outputs":[]},{"cell_type":"code","source":"import pickle \npickle_out = open(\"model.pkl\", mode = \"wb\") \npickle.dump(model, pickle_out) \npickle_out.close()","metadata":{"execution":{"iopub.status.busy":"2022-08-12T09:14:51.793893Z","iopub.execute_input":"2022-08-12T09:14:51.794966Z","iopub.status.idle":"2022-08-12T09:14:51.801378Z","shell.execute_reply.started":"2022-08-12T09:14:51.794912Z","shell.execute_reply":"2022-08-12T09:14:51.800110Z"},"trusted":true},"execution_count":109,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}