Data_Cleaning.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook I have done:\n",
    "1. Exploratory Data Analysis\n",
    "2. Data Cleaning:\n",
    "    1. Selecting relevant features\n",
    "    2. Null value imputation\n",
    "    3. Creating dummy variables\n",
    "    4. Handling outliers\n",
    "    5. Multicolinearity Check"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:24.975672Z",
     "start_time": "2019-01-26T14:22:23.840665Z"
    }
   },
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "from scipy.special import boxcox1p        \n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:24.980637Z",
     "start_time": "2019-01-26T14:22:24.977648Z"
    }
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:25.127590Z",
     "start_time": "2019-01-26T14:22:24.983608Z"
    }
   },
   "outputs": [],
   "source": [
    "pd.set_option('max_colwidth',500)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:34.897349Z",
     "start_time": "2019-01-26T14:22:25.129589Z"
    }
   },
   "outputs": [],
   "source": [
    "data=pd.read_csv('loan.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:34.904322Z",
     "start_time": "2019-01-26T14:22:34.900334Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# data.info()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Dataset has 74 columns including the output columns 'loan_status'.\n",
    "\n",
    "I created a dataframe from data dictionary so that evaluating each column will be easier. This dataframe shows feature name, decription, first value, data type and number of null values."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:35.070622Z",
     "start_time": "2019-01-26T14:22:34.907317Z"
    }
   },
   "outputs": [],
   "source": [
    "data_dict=pd.read_excel('LCDataDictionary.xlsx',sheet_name='LoanStats')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:35.158814Z",
     "start_time": "2019-01-26T14:22:35.072596Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>addr_state</td>\n",
       "      <td>The state provided by the borrower in the loan application</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>annual_inc</td>\n",
       "      <td>The self-reported annual income provided by the borrower during registration.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>annual_inc_joint</td>\n",
       "      <td>The combined self-reported annual income provided by the co-borrowers during registration</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>application_type</td>\n",
       "      <td>Indicates whether the loan is an individual application or a joint application with two co-borrowers</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>collection_recovery_fee</td>\n",
       "      <td>post charge off collection fee</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               LoanStatNew  \\\n",
       "0               addr_state   \n",
       "1               annual_inc   \n",
       "2         annual_inc_joint   \n",
       "3         application_type   \n",
       "4  collection_recovery_fee   \n",
       "\n",
       "                                                                                            Description  \n",
       "0                                            The state provided by the borrower in the loan application  \n",
       "1                         The self-reported annual income provided by the borrower during registration.  \n",
       "2             The combined self-reported annual income provided by the co-borrowers during registration  \n",
       "3  Indicates whether the loan is an individual application or a joint application with two co-borrowers  \n",
       "4                                                                        post charge off collection fee  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dict.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:35.342610Z",
     "start_time": "2019-01-26T14:22:35.160781Z"
    }
   },
   "outputs": [],
   "source": [
    "data_dtypes=pd.DataFrame(data.dtypes,columns=['dtype'])\n",
    "data_dtypes.reset_index(inplace=True)\n",
    "data_dtypes.columns=['LoanStatNew', 'dtype']\n",
    "data_dtypes['first value'] = data.loc[0].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:35.440066Z",
     "start_time": "2019-01-26T14:22:35.343569Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1077501</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>member_id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1296599</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>loan_amnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>5000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>funded_amnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>5000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>funded_amnt_inv</td>\n",
       "      <td>float64</td>\n",
       "      <td>4975</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       LoanStatNew    dtype first value\n",
       "0               id    int64     1077501\n",
       "1        member_id    int64     1296599\n",
       "2        loan_amnt  float64        5000\n",
       "3      funded_amnt  float64        5000\n",
       "4  funded_amnt_inv  float64        4975"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_dtypes.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:35.596147Z",
     "start_time": "2019-01-26T14:22:35.442035Z"
    }
   },
   "outputs": [],
   "source": [
    "data_col=data_dtypes.merge(data_dict, on='LoanStatNew',how='left')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:35.773494Z",
     "start_time": "2019-01-26T14:22:35.598773Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1077501</td>\n",
       "      <td>A unique LC assigned ID for the loan listing.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>member_id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1296599</td>\n",
       "      <td>A unique LC assigned Id for the borrower member.</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  LoanStatNew  dtype first value  \\\n",
       "0          id  int64     1077501   \n",
       "1   member_id  int64     1296599   \n",
       "\n",
       "                                        Description  \n",
       "0     A unique LC assigned ID for the loan listing.  \n",
       "1  A unique LC assigned Id for the borrower member.  "
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col.head(2)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:46.922998Z",
     "start_time": "2019-01-26T14:22:35.775491Z"
    }
   },
   "outputs": [],
   "source": [
    "d={'LoanStatNew':[],'Nulls':[]}\n",
    "for i in data.columns:\n",
    "    d['LoanStatNew'].append(i)\n",
    "    d['Nulls'].append(len(data[data[i].isnull()][i]))\n",
    "data_col=data_col.merge(pd.DataFrame(d),on='LoanStatNew')\n",
    "nulls=pd.DataFrame(d)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:46.937960Z",
     "start_time": "2019-01-26T14:22:46.924993Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "      <th>Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1077501</td>\n",
       "      <td>A unique LC assigned ID for the loan listing.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>member_id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1296599</td>\n",
       "      <td>A unique LC assigned Id for the borrower member.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>loan_amnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>5000</td>\n",
       "      <td>The listed amount of the loan applied for by the borrower. If at some point in time, the credit department reduces the loan amount, then it will be reflected in this value.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>funded_amnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>5000</td>\n",
       "      <td>The total amount committed to that loan at that point in time.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>funded_amnt_inv</td>\n",
       "      <td>float64</td>\n",
       "      <td>4975</td>\n",
       "      <td>The total amount committed by investors for that loan at that point in time.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       LoanStatNew    dtype first value  \\\n",
       "0               id    int64     1077501   \n",
       "1        member_id    int64     1296599   \n",
       "2        loan_amnt  float64        5000   \n",
       "3      funded_amnt  float64        5000   \n",
       "4  funded_amnt_inv  float64        4975   \n",
       "\n",
       "                                                                                                                                                                    Description  \\\n",
       "0                                                                                                                                 A unique LC assigned ID for the loan listing.   \n",
       "1                                                                                                                              A unique LC assigned Id for the borrower member.   \n",
       "2  The listed amount of the loan applied for by the borrower. If at some point in time, the credit department reduces the loan amount, then it will be reflected in this value.   \n",
       "3                                                                                                                The total amount committed to that loan at that point in time.   \n",
       "4                                                                                                  The total amount committed by investors for that loan at that point in time.   \n",
       "\n",
       "   Nulls  \n",
       "0      0  \n",
       "1      0  \n",
       "2      0  \n",
       "3      0  \n",
       "4      0  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now I am able to see info of each column. With this info I decided which column to delete from model."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Exploratory Data Analysis:\n",
    "\n",
    "Here I will go throught important features and analyse them. We'll use this information while data cleaning as well as training machine learning model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:47.189778Z",
     "start_time": "2019-01-26T14:22:46.940954Z"
    }
   },
   "outputs": [],
   "source": [
    "data_y=data['loan_status']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:47.346762Z",
     "start_time": "2019-01-26T14:22:47.192142Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['Fully Paid', 'Charged Off', 'Current', 'Default',\n",
       "       'Late (31-120 days)', 'In Grace Period', 'Late (16-30 days)',\n",
       "       'Does not meet the credit policy. Status:Fully Paid',\n",
       "       'Does not meet the credit policy. Status:Charged Off', 'Issued'],\n",
       "      dtype=object)"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_y.unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:47.663434Z",
     "start_time": "2019-01-26T14:22:47.349723Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Current                                                601779\n",
       "Fully Paid                                             207723\n",
       "Charged Off                                             45248\n",
       "Late (31-120 days)                                      11591\n",
       "Issued                                                   8460\n",
       "In Grace Period                                          6253\n",
       "Late (16-30 days)                                        2357\n",
       "Does not meet the credit policy. Status:Fully Paid       1988\n",
       "Default                                                  1219\n",
       "Does not meet the credit policy. Status:Charged Off       761\n",
       "Name: loan_status, dtype: int64"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_y.value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**I considered Fully Paid loans and Charged off, Late (31-120 days), Default loans as these categories tells surely about loan repayment**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:47.944031Z",
     "start_time": "2019-01-26T14:22:47.665428Z"
    }
   },
   "outputs": [],
   "source": [
    "default=['Charged Off','Late (31-120 days)','Default']\n",
    "data['loan_status']=data['loan_status'].apply(lambda x: 'Default' if x in default else x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:48.041732Z",
     "start_time": "2019-01-26T14:22:47.946006Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Current                                                601779\n",
       "Fully Paid                                             207723\n",
       "Default                                                 58058\n",
       "Issued                                                   8460\n",
       "In Grace Period                                          6253\n",
       "Late (16-30 days)                                        2357\n",
       "Does not meet the credit policy. Status:Fully Paid       1988\n",
       "Does not meet the credit policy. Status:Charged Off       761\n",
       "Name: loan_status, dtype: int64"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_y.value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:48.415050Z",
     "start_time": "2019-01-26T14:22:48.043719Z"
    }
   },
   "outputs": [],
   "source": [
    "data=data[(data['loan_status']=='Default') | (data['loan_status']=='Fully Paid')]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Following are some important points about data:\n",
    "1. Data is skewed with the ratio of **22/78**. That is, in 100 rows, only 22 rows refers to 'Default' case which is expected as defaulted cases will always be minimal.\n",
    "2. Loan defaulters tend to apply for larger loans.\n",
    "3. Around 23% of funded amount is defaulted"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:48.449984Z",
     "start_time": "2019-01-26T14:22:48.417046Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.7815569961735414"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Data is skewed with the ration of 21.84/78.16\n",
    "data['loan_status'].value_counts()[0]/len(data['loan_status'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:48.585650Z",
     "start_time": "2019-01-26T14:22:48.450954Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>loan_amnt</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>loan_status</th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>Default</th>\n",
       "      <td>14771.922905</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>Fully Paid</th>\n",
       "      <td>13346.350910</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                loan_amnt\n",
       "loan_status              \n",
       "Default      14771.922905\n",
       "Fully Paid   13346.350910"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#relation between average loan amount and loan status\n",
    "data[['loan_amnt','loan_status']].groupby(by='loan_status').mean()\n",
    "\n",
    "#Loan defaulters tend to apply for larger loans."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:49.200919Z",
     "start_time": "2019-01-26T14:22:48.588643Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2851b917860>"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,6))\n",
    "sns.boxplot(data['loan_amnt'],data['loan_status'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:49.206903Z",
     "start_time": "2019-01-26T14:22:49.202913Z"
    }
   },
   "outputs": [],
   "source": [
    "# data.to_csv('DC1_1.csv',index=False)\n",
    "# data=pd.read_csv('DC1_1.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:49.535535Z",
     "start_time": "2019-01-26T14:22:49.209896Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2850dd92eb8>"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "total_funded_amnt=data['funded_amnt'].sum()\n",
    "(data[['funded_amnt','loan_status']].groupby(by='loan_status').\\\n",
    "         sum()/total_funded_amnt*100).plot(kind='bar')\n",
    "\n",
    "#Around 23% of funded amount is defaulted"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Now see how loan default rate and interest rate changes with grade and year.\n",
    "I did following aggregations:\n",
    "1. Default rate by grade\n",
    "2. Default rate by year\n",
    "3. Interest rate by grade\n",
    "4. Interest rate by year\n",
    "\n",
    "Following are the findings:\n",
    "1. Default rate highly correlated with grade. Riskier the grade more is the chance to default\n",
    "2. Defalut rate do not show any pattern with issue year, but it seems to relate with the economic conditions prevailing at that time. Also after 2010 the loans issued on site increased which also increased the loan defaulters. 2014 has peak default rate.\n",
    "3. Interest rate is highly correlated with grade. Riskier the grade more is the interest rate.\n",
    "4. Average interest rate is increasing. It is at peak in 2013 and 2014 which may be a reason for peak default rate in 2014 "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:49.560422Z",
     "start_time": "2019-01-26T14:22:49.537531Z"
    }
   },
   "outputs": [],
   "source": [
    "def get_default_rate_by_col(column):\n",
    "    \"\"\"\n",
    "    Param:\n",
    "        column: str, column for which default rate is to be calculated for it's unique values. Column must have limited\n",
    "        unique values.\n",
    "    \"\"\"\n",
    "    tmp=pd.DataFrame(data[data['loan_status']=='Default'][column].value_counts()/data[column].value_counts())\n",
    "    tmp.columns=[column+'_Default_Rate']\n",
    "    tmp=tmp.join(data[column].value_counts())\n",
    "    tmp.columns=[column+'_Default_Rate',column+'_Count']\n",
    "    return tmp"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:49.911295Z",
     "start_time": "2019-01-26T14:22:49.561385Z"
    }
   },
   "outputs": [],
   "source": [
    "# Create a column issue_year which has year in which the loan was issued.\n",
    "data['issue_year']=data['issue_d'].apply(lambda x:int(x.split('-')[1]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:50.133252Z",
     "start_time": "2019-01-26T14:22:49.913839Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Default Rate Vs. Grade')"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "default_rate_by_grade=get_default_rate_by_col('grade')\n",
    "plt.bar(x=default_rate_by_grade.index,height=default_rate_by_grade['grade_Default_Rate'])\n",
    "plt.title('Default Rate Vs. Grade')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:50.309794Z",
     "start_time": "2019-01-26T14:22:50.135248Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Default Rate Vs. Issue Year')"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "default_rate_by_year=get_default_rate_by_col('issue_year')\n",
    "plt.plot(default_rate_by_year['issue_year_Default_Rate'].sort_index())\n",
    "plt.title('Default Rate Vs. Issue Year')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:50.426494Z",
     "start_time": "2019-01-26T14:22:50.311775Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Interest Rate Vs. Grade')"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "interest_by_grade=data[['int_rate','grade']].groupby('grade').mean()\n",
    "plt.bar(x=interest_by_grade.index,height=interest_by_grade['int_rate'])\n",
    "plt.title('Interest Rate Vs. Grade')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:50.572342Z",
     "start_time": "2019-01-26T14:22:50.428474Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Text(0.5,1,'Interest Rate Vs. Year')"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "interest_by_year=data[['int_rate','issue_year']].groupby('issue_year').mean()\n",
    "plt.plot(interest_by_year.sort_index())\n",
    "plt.title('Interest Rate Vs. Year')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Look at the ** Return on Investment **\n",
    "\n",
    "ROI = total_pymnt (amount returned by borrower) / funded_amnt\n",
    "\n",
    "1. Total ROI is in negative as we are considering defaulted records also. As defaulted records generally have higher loan amounts, it moves overall ROI to negative side.\n",
    "2. For riskier grades, overall ROI (incl. defaulted loans) decreases.\n",
    "3. If we consider fully paid loans, then ROI is more for riskier grades."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:50.614253Z",
     "start_time": "2019-01-26T14:22:50.575333Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROI: -2.65\n"
     ]
    }
   ],
   "source": [
    "tmp=data[['total_pymnt','funded_amnt','loan_status','grade']]\n",
    "roi=(tmp['total_pymnt'].sum()/tmp['funded_amnt'].sum()-1)*100\n",
    "print(\"ROI:\",round(roi,2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:50.765178Z",
     "start_time": "2019-01-26T14:22:50.617220Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "grade\n",
       "A    0.034525\n",
       "B    0.027576\n",
       "C   -0.024436\n",
       "D   -0.071834\n",
       "E   -0.115351\n",
       "F   -0.137799\n",
       "G   -0.165548\n",
       "dtype: float64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "tmp1=tmp.groupby('grade').sum()\n",
    "ROI_by_grade=tmp1['total_pymnt']/tmp1['funded_amnt'] -1\n",
    "ROI_by_grade"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:51.019854Z",
     "start_time": "2019-01-26T14:22:50.767171Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROI for fully paid loans: 14.13\n"
     ]
    }
   ],
   "source": [
    "tmp=data[data['loan_status']=='Fully Paid'][['total_pymnt','funded_amnt','loan_status','grade']]\n",
    "roi=(tmp['total_pymnt'].sum()/tmp['funded_amnt'].sum()-1)*100\n",
    "print(\"ROI for fully paid loans:\",round(roi,2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:51.043781Z",
     "start_time": "2019-01-26T14:22:51.021840Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "grade\n",
       "A    0.079777\n",
       "B    0.125364\n",
       "C    0.149117\n",
       "D    0.175479\n",
       "E    0.197441\n",
       "F    0.228829\n",
       "G    0.238275\n",
       "dtype: float64"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Gradewise ROI for fully paid loans\n",
    "tmp1=tmp.groupby('grade').sum()\n",
    "fully_paid_ROI_by_grade=tmp1['total_pymnt']/tmp1['funded_amnt'] -1\n",
    "fully_paid_ROI_by_grade"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning 1:\n",
    "This is initial phase of data cleaning where I am dropping columns by looking at the their decription and the number of null values. As we want to predict loan_status when applicant applies for loan, I have deleted columns which leak information about future.\n",
    "e.g. 'funded_amt', 'last_pymnt_amnt', 'installment', etc."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:51.183861Z",
     "start_time": "2019-01-26T14:22:51.045783Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "      <th>Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1077501</td>\n",
       "      <td>A unique LC assigned ID for the loan listing.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>member_id</td>\n",
       "      <td>int64</td>\n",
       "      <td>1296599</td>\n",
       "      <td>A unique LC assigned Id for the borrower member.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>loan_amnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>5000</td>\n",
       "      <td>The listed amount of the loan applied for by the borrower. If at some point in time, the credit department reduces the loan amount, then it will be reflected in this value.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>funded_amnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>5000</td>\n",
       "      <td>The total amount committed to that loan at that point in time.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>funded_amnt_inv</td>\n",
       "      <td>float64</td>\n",
       "      <td>4975</td>\n",
       "      <td>The total amount committed by investors for that loan at that point in time.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>term</td>\n",
       "      <td>object</td>\n",
       "      <td>36 months</td>\n",
       "      <td>The number of payments on the loan. Values are in months and can be either 36 or 60.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>int_rate</td>\n",
       "      <td>float64</td>\n",
       "      <td>10.65</td>\n",
       "      <td>Interest Rate on the loan</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>installment</td>\n",
       "      <td>float64</td>\n",
       "      <td>162.87</td>\n",
       "      <td>The monthly payment owed by the borrower if the loan originates.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>grade</td>\n",
       "      <td>object</td>\n",
       "      <td>B</td>\n",
       "      <td>LC assigned loan grade</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>sub_grade</td>\n",
       "      <td>object</td>\n",
       "      <td>B2</td>\n",
       "      <td>LC assigned loan subgrade</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>emp_title</td>\n",
       "      <td>object</td>\n",
       "      <td>NaN</td>\n",
       "      <td>The job title supplied by the Borrower when applying for the loan.*</td>\n",
       "      <td>51462</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>emp_length</td>\n",
       "      <td>object</td>\n",
       "      <td>10+ years</td>\n",
       "      <td>Employment length in years. Possible values are between 0 and 10 where 0 means less than one year and 10 means ten or more years.</td>\n",
       "      <td>44825</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>home_ownership</td>\n",
       "      <td>object</td>\n",
       "      <td>RENT</td>\n",
       "      <td>The home ownership status provided by the borrower during registration. Our values are: RENT, OWN, MORTGAGE, OTHER.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        LoanStatNew    dtype first value  \\\n",
       "0                id    int64     1077501   \n",
       "1         member_id    int64     1296599   \n",
       "2         loan_amnt  float64        5000   \n",
       "3       funded_amnt  float64        5000   \n",
       "4   funded_amnt_inv  float64        4975   \n",
       "5              term   object   36 months   \n",
       "6          int_rate  float64       10.65   \n",
       "7       installment  float64      162.87   \n",
       "8             grade   object           B   \n",
       "9         sub_grade   object          B2   \n",
       "10        emp_title   object         NaN   \n",
       "11       emp_length   object   10+ years   \n",
       "12   home_ownership   object        RENT   \n",
       "\n",
       "                                                                                                                                                                     Description  \\\n",
       "0                                                                                                                                  A unique LC assigned ID for the loan listing.   \n",
       "1                                                                                                                               A unique LC assigned Id for the borrower member.   \n",
       "2   The listed amount of the loan applied for by the borrower. If at some point in time, the credit department reduces the loan amount, then it will be reflected in this value.   \n",
       "3                                                                                                                 The total amount committed to that loan at that point in time.   \n",
       "4                                                                                                   The total amount committed by investors for that loan at that point in time.   \n",
       "5                                                                                           The number of payments on the loan. Values are in months and can be either 36 or 60.   \n",
       "6                                                                                                                                                      Interest Rate on the loan   \n",
       "7                                                                                                               The monthly payment owed by the borrower if the loan originates.   \n",
       "8                                                                                                                                                         LC assigned loan grade   \n",
       "9                                                                                                                                                      LC assigned loan subgrade   \n",
       "10                                                                                                           The job title supplied by the Borrower when applying for the loan.*   \n",
       "11                                            Employment length in years. Possible values are between 0 and 10 where 0 means less than one year and 10 means ten or more years.    \n",
       "12                                                           The home ownership status provided by the borrower during registration. Our values are: RENT, OWN, MORTGAGE, OTHER.   \n",
       "\n",
       "    Nulls  \n",
       "0       0  \n",
       "1       0  \n",
       "2       0  \n",
       "3       0  \n",
       "4       0  \n",
       "5       0  \n",
       "6       0  \n",
       "7       0  \n",
       "8       0  \n",
       "9       0  \n",
       "10  51462  \n",
       "11  44825  \n",
       "12      0  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col[0:13]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:51.478520Z",
     "start_time": "2019-01-26T14:22:51.185830Z"
    }
   },
   "outputs": [],
   "source": [
    "#'grade' and 'sub_grade' gives same information\n",
    "data.drop(['id','member_id','funded_amnt','funded_amnt_inv','installment','sub_grade','emp_title'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:51.605350Z",
     "start_time": "2019-01-26T14:22:51.481512Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "      <th>Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>annual_inc</td>\n",
       "      <td>float64</td>\n",
       "      <td>24000</td>\n",
       "      <td>The self-reported annual income provided by the borrower during registration.</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>verification_status</td>\n",
       "      <td>object</td>\n",
       "      <td>Verified</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>issue_d</td>\n",
       "      <td>object</td>\n",
       "      <td>Dec-2011</td>\n",
       "      <td>The month which the loan was funded</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>loan_status</td>\n",
       "      <td>object</td>\n",
       "      <td>Fully Paid</td>\n",
       "      <td>Current status of the loan</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>pymnt_plan</td>\n",
       "      <td>object</td>\n",
       "      <td>n</td>\n",
       "      <td>Indicates if a payment plan has been put in place for the loan</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>url</td>\n",
       "      <td>object</td>\n",
       "      <td>https://www.lendingclub.com/browse/loanDetail.action?loan_id=1077501</td>\n",
       "      <td>URL for the LC page with listing data.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>desc</td>\n",
       "      <td>object</td>\n",
       "      <td>Borrower added on 12/22/11 &gt; I need to upgrade my business technologies.&lt;br&gt;</td>\n",
       "      <td>Loan description provided by the borrower</td>\n",
       "      <td>761351</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>purpose</td>\n",
       "      <td>object</td>\n",
       "      <td>credit_card</td>\n",
       "      <td>A category provided by the borrower for the loan request.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>title</td>\n",
       "      <td>object</td>\n",
       "      <td>Computer</td>\n",
       "      <td>The loan title provided by the borrower</td>\n",
       "      <td>152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>zip_code</td>\n",
       "      <td>object</td>\n",
       "      <td>860xx</td>\n",
       "      <td>The first 3 numbers of the zip code provided by the borrower in the loan application.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>addr_state</td>\n",
       "      <td>object</td>\n",
       "      <td>AZ</td>\n",
       "      <td>The state provided by the borrower in the loan application</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>dti</td>\n",
       "      <td>float64</td>\n",
       "      <td>27.65</td>\n",
       "      <td>A ratio calculated using the borrower’s total monthly debt payments on the total debt obligations, excluding mortgage and the requested LC loan, divided by the borrower’s self-reported monthly income.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>delinq_2yrs</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>The number of 30+ days past-due incidences of delinquency in the borrower's credit file for the past 2 years</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "            LoanStatNew    dtype  \\\n",
       "13           annual_inc  float64   \n",
       "14  verification_status   object   \n",
       "15              issue_d   object   \n",
       "16          loan_status   object   \n",
       "17           pymnt_plan   object   \n",
       "18                  url   object   \n",
       "19                 desc   object   \n",
       "20              purpose   object   \n",
       "21                title   object   \n",
       "22             zip_code   object   \n",
       "23           addr_state   object   \n",
       "24                  dti  float64   \n",
       "25          delinq_2yrs  float64   \n",
       "\n",
       "                                                                       first value  \\\n",
       "13                                                                           24000   \n",
       "14                                                                        Verified   \n",
       "15                                                                        Dec-2011   \n",
       "16                                                                      Fully Paid   \n",
       "17                                                                               n   \n",
       "18            https://www.lendingclub.com/browse/loanDetail.action?loan_id=1077501   \n",
       "19    Borrower added on 12/22/11 > I need to upgrade my business technologies.<br>   \n",
       "20                                                                     credit_card   \n",
       "21                                                                        Computer   \n",
       "22                                                                           860xx   \n",
       "23                                                                              AZ   \n",
       "24                                                                           27.65   \n",
       "25                                                                               0   \n",
       "\n",
       "                                                                                                                                                                                                 Description  \\\n",
       "13                                                                                                                             The self-reported annual income provided by the borrower during registration.   \n",
       "14                                                                                                                                                                                                       NaN   \n",
       "15                                                                                                                                                                       The month which the loan was funded   \n",
       "16                                                                                                                                                                                Current status of the loan   \n",
       "17                                                                                                                                            Indicates if a payment plan has been put in place for the loan   \n",
       "18                                                                                                                                                                    URL for the LC page with listing data.   \n",
       "19                                                                                                                                                                 Loan description provided by the borrower   \n",
       "20                                                                                                                                                A category provided by the borrower for the loan request.    \n",
       "21                                                                                                                                                                   The loan title provided by the borrower   \n",
       "22                                                                                                                     The first 3 numbers of the zip code provided by the borrower in the loan application.   \n",
       "23                                                                                                                                                The state provided by the borrower in the loan application   \n",
       "24  A ratio calculated using the borrower’s total monthly debt payments on the total debt obligations, excluding mortgage and the requested LC loan, divided by the borrower’s self-reported monthly income.   \n",
       "25                                                                                              The number of 30+ days past-due incidences of delinquency in the borrower's credit file for the past 2 years   \n",
       "\n",
       "     Nulls  \n",
       "13       4  \n",
       "14       0  \n",
       "15       0  \n",
       "16       0  \n",
       "17       0  \n",
       "18       0  \n",
       "19  761351  \n",
       "20       0  \n",
       "21     152  \n",
       "22       0  \n",
       "23       0  \n",
       "24       0  \n",
       "25      29  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col[13:26]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:51.898558Z",
     "start_time": "2019-01-26T14:22:51.608333Z"
    }
   },
   "outputs": [],
   "source": [
    "data.drop(['url','desc','title','zip_code'],axis=1,inplace=True)\n",
    "#Info in 'title' and 'purpose' is same"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:52.010259Z",
     "start_time": "2019-01-26T14:22:51.902549Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "      <th>Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>earliest_cr_line</td>\n",
       "      <td>object</td>\n",
       "      <td>Jan-1985</td>\n",
       "      <td>The month the borrower's earliest reported credit line was opened</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>inq_last_6mths</td>\n",
       "      <td>float64</td>\n",
       "      <td>1</td>\n",
       "      <td>The number of inquiries in past 6 months (excluding auto and mortgage inquiries)</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>mths_since_last_delinq</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>The number of months since the borrower's last delinquency.</td>\n",
       "      <td>454312</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>mths_since_last_record</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>The number of months since the last public record.</td>\n",
       "      <td>750326</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>30</th>\n",
       "      <td>open_acc</td>\n",
       "      <td>float64</td>\n",
       "      <td>3</td>\n",
       "      <td>The number of open credit lines in the borrower's credit file.</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>31</th>\n",
       "      <td>pub_rec</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>Number of derogatory public records</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>32</th>\n",
       "      <td>revol_bal</td>\n",
       "      <td>float64</td>\n",
       "      <td>13648</td>\n",
       "      <td>Total credit revolving balance</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>33</th>\n",
       "      <td>revol_util</td>\n",
       "      <td>float64</td>\n",
       "      <td>83.7</td>\n",
       "      <td>Revolving line utilization rate, or the amount of credit the borrower is using relative to all available revolving credit.</td>\n",
       "      <td>502</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>34</th>\n",
       "      <td>total_acc</td>\n",
       "      <td>float64</td>\n",
       "      <td>9</td>\n",
       "      <td>The total number of credit lines currently in the borrower's credit file</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35</th>\n",
       "      <td>initial_list_status</td>\n",
       "      <td>object</td>\n",
       "      <td>f</td>\n",
       "      <td>The initial listing status of the loan. Possible values are – W, F</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36</th>\n",
       "      <td>out_prncp</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>Remaining outstanding principal for total amount funded</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>37</th>\n",
       "      <td>out_prncp_inv</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>Remaining outstanding principal for portion of total amount funded by investors</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38</th>\n",
       "      <td>total_pymnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>5861.07</td>\n",
       "      <td>Payments received to date for total amount funded</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               LoanStatNew    dtype first value  \\\n",
       "26        earliest_cr_line   object    Jan-1985   \n",
       "27          inq_last_6mths  float64           1   \n",
       "28  mths_since_last_delinq  float64         NaN   \n",
       "29  mths_since_last_record  float64         NaN   \n",
       "30                open_acc  float64           3   \n",
       "31                 pub_rec  float64           0   \n",
       "32               revol_bal  float64       13648   \n",
       "33              revol_util  float64        83.7   \n",
       "34               total_acc  float64           9   \n",
       "35     initial_list_status   object           f   \n",
       "36               out_prncp  float64           0   \n",
       "37           out_prncp_inv  float64           0   \n",
       "38             total_pymnt  float64     5861.07   \n",
       "\n",
       "                                                                                                                   Description  \\\n",
       "26                                                           The month the borrower's earliest reported credit line was opened   \n",
       "27                                            The number of inquiries in past 6 months (excluding auto and mortgage inquiries)   \n",
       "28                                                                 The number of months since the borrower's last delinquency.   \n",
       "29                                                                          The number of months since the last public record.   \n",
       "30                                                              The number of open credit lines in the borrower's credit file.   \n",
       "31                                                                                         Number of derogatory public records   \n",
       "32                                                                                              Total credit revolving balance   \n",
       "33  Revolving line utilization rate, or the amount of credit the borrower is using relative to all available revolving credit.   \n",
       "34                                                    The total number of credit lines currently in the borrower's credit file   \n",
       "35                                                          The initial listing status of the loan. Possible values are – W, F   \n",
       "36                                                                     Remaining outstanding principal for total amount funded   \n",
       "37                                             Remaining outstanding principal for portion of total amount funded by investors   \n",
       "38                                                                           Payments received to date for total amount funded   \n",
       "\n",
       "     Nulls  \n",
       "26      29  \n",
       "27      29  \n",
       "28  454312  \n",
       "29  750326  \n",
       "30      29  \n",
       "31      29  \n",
       "32       0  \n",
       "33     502  \n",
       "34      29  \n",
       "35       0  \n",
       "36       0  \n",
       "37       0  \n",
       "38       0  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col[26:39]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:52.313446Z",
     "start_time": "2019-01-26T14:22:52.013251Z"
    }
   },
   "outputs": [],
   "source": [
    "data.drop(['mths_since_last_delinq','mths_since_last_record','out_prncp','out_prncp_inv','total_pymnt'],\n",
    "          axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:52.376288Z",
     "start_time": "2019-01-26T14:22:52.316447Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "      <th>Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>39</th>\n",
       "      <td>total_pymnt_inv</td>\n",
       "      <td>float64</td>\n",
       "      <td>5831.78</td>\n",
       "      <td>Payments received to date for portion of total amount funded by investors</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>40</th>\n",
       "      <td>total_rec_prncp</td>\n",
       "      <td>float64</td>\n",
       "      <td>5000</td>\n",
       "      <td>Principal received to date</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41</th>\n",
       "      <td>total_rec_int</td>\n",
       "      <td>float64</td>\n",
       "      <td>861.07</td>\n",
       "      <td>Interest received to date</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42</th>\n",
       "      <td>total_rec_late_fee</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>Late fees received to date</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>43</th>\n",
       "      <td>recoveries</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>post charge off gross recovery</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44</th>\n",
       "      <td>collection_recovery_fee</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>post charge off collection fee</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45</th>\n",
       "      <td>last_pymnt_d</td>\n",
       "      <td>object</td>\n",
       "      <td>Jan-2015</td>\n",
       "      <td>Last month payment was received</td>\n",
       "      <td>17659</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>46</th>\n",
       "      <td>last_pymnt_amnt</td>\n",
       "      <td>float64</td>\n",
       "      <td>171.62</td>\n",
       "      <td>Last total payment amount received</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>47</th>\n",
       "      <td>next_pymnt_d</td>\n",
       "      <td>object</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Next scheduled payment date</td>\n",
       "      <td>252971</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>48</th>\n",
       "      <td>last_credit_pull_d</td>\n",
       "      <td>object</td>\n",
       "      <td>Jan-2016</td>\n",
       "      <td>The most recent month LC pulled credit for this loan</td>\n",
       "      <td>53</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>49</th>\n",
       "      <td>collections_12_mths_ex_med</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>Number of collections in 12 months excluding medical collections</td>\n",
       "      <td>145</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>50</th>\n",
       "      <td>mths_since_last_major_derog</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Months since most recent 90-day or worse rating</td>\n",
       "      <td>665676</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>51</th>\n",
       "      <td>policy_code</td>\n",
       "      <td>float64</td>\n",
       "      <td>1</td>\n",
       "      <td>publicly available policy_code=1\\nnew products not publicly available policy_code=2</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    LoanStatNew    dtype first value  \\\n",
       "39              total_pymnt_inv  float64     5831.78   \n",
       "40              total_rec_prncp  float64        5000   \n",
       "41                total_rec_int  float64      861.07   \n",
       "42           total_rec_late_fee  float64           0   \n",
       "43                   recoveries  float64           0   \n",
       "44      collection_recovery_fee  float64           0   \n",
       "45                 last_pymnt_d   object    Jan-2015   \n",
       "46              last_pymnt_amnt  float64      171.62   \n",
       "47                 next_pymnt_d   object         NaN   \n",
       "48           last_credit_pull_d   object    Jan-2016   \n",
       "49   collections_12_mths_ex_med  float64           0   \n",
       "50  mths_since_last_major_derog  float64         NaN   \n",
       "51                  policy_code  float64           1   \n",
       "\n",
       "                                                                            Description  \\\n",
       "39            Payments received to date for portion of total amount funded by investors   \n",
       "40                                                           Principal received to date   \n",
       "41                                                            Interest received to date   \n",
       "42                                                           Late fees received to date   \n",
       "43                                                       post charge off gross recovery   \n",
       "44                                                       post charge off collection fee   \n",
       "45                                                      Last month payment was received   \n",
       "46                                                   Last total payment amount received   \n",
       "47                                                          Next scheduled payment date   \n",
       "48                                 The most recent month LC pulled credit for this loan   \n",
       "49                     Number of collections in 12 months excluding medical collections   \n",
       "50                                      Months since most recent 90-day or worse rating   \n",
       "51  publicly available policy_code=1\\nnew products not publicly available policy_code=2   \n",
       "\n",
       "     Nulls  \n",
       "39       0  \n",
       "40       0  \n",
       "41       0  \n",
       "42       0  \n",
       "43       0  \n",
       "44       0  \n",
       "45   17659  \n",
       "46       0  \n",
       "47  252971  \n",
       "48      53  \n",
       "49     145  \n",
       "50  665676  \n",
       "51       0  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col[39:52]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:52.646559Z",
     "start_time": "2019-01-26T14:22:52.378273Z"
    }
   },
   "outputs": [],
   "source": [
    "data.drop(['total_pymnt_inv','total_rec_prncp','total_rec_int','total_rec_late_fee','recoveries',\n",
    "          'collection_recovery_fee','last_pymnt_d','last_pymnt_amnt','next_pymnt_d','last_credit_pull_d','policy_code',\n",
    "          'mths_since_last_major_derog'],\n",
    "          axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:52.691436Z",
     "start_time": "2019-01-26T14:22:52.656532Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "      <th>Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>52</th>\n",
       "      <td>application_type</td>\n",
       "      <td>object</td>\n",
       "      <td>INDIVIDUAL</td>\n",
       "      <td>Indicates whether the loan is an individual application or a joint application with two co-borrowers</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>53</th>\n",
       "      <td>annual_inc_joint</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>The combined self-reported annual income provided by the co-borrowers during registration</td>\n",
       "      <td>886868</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>54</th>\n",
       "      <td>dti_joint</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>A ratio calculated using the co-borrowers' total monthly payments on the total debt obligations, excluding mortgages and the requested LC loan, divided by the co-borrowers' combined self-reported monthly income</td>\n",
       "      <td>886870</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>55</th>\n",
       "      <td>verification_status_joint</td>\n",
       "      <td>object</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>886868</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>56</th>\n",
       "      <td>acc_now_delinq</td>\n",
       "      <td>float64</td>\n",
       "      <td>0</td>\n",
       "      <td>The number of accounts on which the borrower is now delinquent.</td>\n",
       "      <td>29</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>57</th>\n",
       "      <td>tot_coll_amt</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Total collection amounts ever owed</td>\n",
       "      <td>70276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>58</th>\n",
       "      <td>tot_cur_bal</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Total current balance of all accounts</td>\n",
       "      <td>70276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>59</th>\n",
       "      <td>open_acc_6m</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of open trades in last 6 months</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>60</th>\n",
       "      <td>open_il_6m</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of currently active installment trades</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>61</th>\n",
       "      <td>open_il_12m</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of installment accounts opened in past 12 months</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>62</th>\n",
       "      <td>open_il_24m</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of installment accounts opened in past 24 months</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>63</th>\n",
       "      <td>mths_since_rcnt_il</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Months since most recent installment accounts opened</td>\n",
       "      <td>866569</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>64</th>\n",
       "      <td>total_bal_il</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Total current balance of all installment accounts</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  LoanStatNew    dtype first value  \\\n",
       "52           application_type   object  INDIVIDUAL   \n",
       "53           annual_inc_joint  float64         NaN   \n",
       "54                  dti_joint  float64         NaN   \n",
       "55  verification_status_joint   object         NaN   \n",
       "56             acc_now_delinq  float64           0   \n",
       "57               tot_coll_amt  float64         NaN   \n",
       "58                tot_cur_bal  float64         NaN   \n",
       "59                open_acc_6m  float64         NaN   \n",
       "60                 open_il_6m  float64         NaN   \n",
       "61                open_il_12m  float64         NaN   \n",
       "62                open_il_24m  float64         NaN   \n",
       "63         mths_since_rcnt_il  float64         NaN   \n",
       "64               total_bal_il  float64         NaN   \n",
       "\n",
       "                                                                                                                                                                                                           Description  \\\n",
       "52                                                                                                                Indicates whether the loan is an individual application or a joint application with two co-borrowers   \n",
       "53                                                                                                                           The combined self-reported annual income provided by the co-borrowers during registration   \n",
       "54  A ratio calculated using the co-borrowers' total monthly payments on the total debt obligations, excluding mortgages and the requested LC loan, divided by the co-borrowers' combined self-reported monthly income   \n",
       "55                                                                                                                                                                                                                 NaN   \n",
       "56                                                                                                                                                     The number of accounts on which the borrower is now delinquent.   \n",
       "57                                                                                                                                                                                  Total collection amounts ever owed   \n",
       "58                                                                                                                                                                               Total current balance of all accounts   \n",
       "59                                                                                                                                                                              Number of open trades in last 6 months   \n",
       "60                                                                                                                                                                       Number of currently active installment trades   \n",
       "61                                                                                                                                                             Number of installment accounts opened in past 12 months   \n",
       "62                                                                                                                                                             Number of installment accounts opened in past 24 months   \n",
       "63                                                                                                                                                                Months since most recent installment accounts opened   \n",
       "64                                                                                                                                                                   Total current balance of all installment accounts   \n",
       "\n",
       "     Nulls  \n",
       "52       0  \n",
       "53  886868  \n",
       "54  886870  \n",
       "55  886868  \n",
       "56      29  \n",
       "57   70276  \n",
       "58   70276  \n",
       "59  866007  \n",
       "60  866007  \n",
       "61  866007  \n",
       "62  866007  \n",
       "63  866569  \n",
       "64  866007  "
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col[52:65]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:53.307788Z",
     "start_time": "2019-01-26T14:22:52.694429Z"
    }
   },
   "outputs": [],
   "source": [
    "# Drop any column with more than 50% missing values\n",
    "data = data.dropna(thresh=(len(data)/2),axis=1) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:53.331730Z",
     "start_time": "2019-01-26T14:22:53.310780Z"
    }
   },
   "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>LoanStatNew</th>\n",
       "      <th>dtype</th>\n",
       "      <th>first value</th>\n",
       "      <th>Description</th>\n",
       "      <th>Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>65</th>\n",
       "      <td>il_util</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Ratio of total current balance to high credit/credit limit on all install acct</td>\n",
       "      <td>868762</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>66</th>\n",
       "      <td>open_rv_12m</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of revolving trades opened in past 12 months</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>67</th>\n",
       "      <td>open_rv_24m</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of revolving trades opened in past 24 months</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>68</th>\n",
       "      <td>max_bal_bc</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Maximum current balance owed on all revolving accounts</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>all_util</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Balance to credit limit on all trades</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>70</th>\n",
       "      <td>total_rev_hi_lim</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>70276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>71</th>\n",
       "      <td>inq_fi</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of personal finance inquiries</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>72</th>\n",
       "      <td>total_cu_tl</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of finance trades</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>73</th>\n",
       "      <td>inq_last_12m</td>\n",
       "      <td>float64</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Number of credit inquiries in past 12 months</td>\n",
       "      <td>866007</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "         LoanStatNew    dtype first value  \\\n",
       "65           il_util  float64         NaN   \n",
       "66       open_rv_12m  float64         NaN   \n",
       "67       open_rv_24m  float64         NaN   \n",
       "68        max_bal_bc  float64         NaN   \n",
       "69          all_util  float64         NaN   \n",
       "70  total_rev_hi_lim  float64         NaN   \n",
       "71            inq_fi  float64         NaN   \n",
       "72       total_cu_tl  float64         NaN   \n",
       "73      inq_last_12m  float64         NaN   \n",
       "\n",
       "                                                                       Description  \\\n",
       "65  Ratio of total current balance to high credit/credit limit on all install acct   \n",
       "66                             Number of revolving trades opened in past 12 months   \n",
       "67                             Number of revolving trades opened in past 24 months   \n",
       "68                          Maximum current balance owed on all revolving accounts   \n",
       "69                                           Balance to credit limit on all trades   \n",
       "70                                                                             NaN   \n",
       "71                                            Number of personal finance inquiries   \n",
       "72                                                        Number of finance trades   \n",
       "73                                    Number of credit inquiries in past 12 months   \n",
       "\n",
       "     Nulls  \n",
       "65  868762  \n",
       "66  866007  \n",
       "67  866007  \n",
       "68  866007  \n",
       "69  866007  \n",
       "70   70276  \n",
       "71  866007  \n",
       "72  866007  \n",
       "73  866007  "
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data_col[65:]\n",
    "#These columns are already removed as they have more than 50% null values (except total_hi_rev_lim)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:53.443067Z",
     "start_time": "2019-01-26T14:22:53.335713Z"
    }
   },
   "outputs": [],
   "source": [
    "# data.to_csv('data_dc1.csv',index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning 2 : Null Imputation\n",
    "After initial data cleaning, we're left with 30 columns. Now I did analysis of each column"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:53.584839Z",
     "start_time": "2019-01-26T14:22:53.445064Z"
    }
   },
   "outputs": [],
   "source": [
    "# data=pd.read_csv('data_dc1.csv')\n",
    "# data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "I calculated percent null values in each column.\n",
    "1. Columns with null values : 'emp_length', 'revol_util', 'collections_12_mths_ex_med', 'tot_coll_amt', 'tot_cur_bal', 'total_rev_hi_lim'\n",
    "2. As 'emp_length' is ordinal variable, I encoded it according to value instead of creating dummy binary variables. 'emp_length' less that one year was set to 0 and that greater than 10 years was set to 10.\n",
    "3. 'tot_coll_amt','tot_cur_bal','total_rev_hi_lim' have missing values till Sept-2012 date.\n",
    "4. Other variables do not show any pattern in missing values\n",
    "\n",
    "Check histograms of these columns. All of them are skewed with wide outliers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:53.975104Z",
     "start_time": "2019-01-26T14:22:53.587832Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>% Nulls</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>emp_length</th>\n",
       "      <td>4.02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>revol_util</th>\n",
       "      <td>0.08</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>collections_12_mths_ex_med</th>\n",
       "      <td>0.02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tot_coll_amt</th>\n",
       "      <td>24.02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tot_cur_bal</th>\n",
       "      <td>24.02</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>total_rev_hi_lim</th>\n",
       "      <td>24.02</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                            % Nulls\n",
       "emp_length                     4.02\n",
       "revol_util                     0.08\n",
       "collections_12_mths_ex_med     0.02\n",
       "tot_coll_amt                  24.02\n",
       "tot_cur_bal                   24.02\n",
       "total_rev_hi_lim              24.02"
      ]
     },
     "execution_count": 48,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "nulls_percent=[]\n",
    "for i in data.columns:\n",
    "    tmp=len(data[data[i].isnull()])/len(data)*100\n",
    "    nulls_percent.append(round(tmp,2))\n",
    "null_df=pd.DataFrame(data=nulls_percent,index=data.columns,columns=['% Nulls'])\n",
    "null_df[null_df['% Nulls']!=0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:53.981094Z",
     "start_time": "2019-01-26T14:22:53.977099Z"
    }
   },
   "outputs": [],
   "source": [
    "columns_with_nulls=['emp_length','revol_util','collections_12_mths_ex_med',\\\n",
    "                    'tot_coll_amt','tot_cur_bal','total_rev_hi_lim']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:54.151631Z",
     "start_time": "2019-01-26T14:22:53.990063Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['10+ years', '< 1 year', '3 years', '9 years', '4 years',\n",
       "       '5 years', '1 year', '6 years', '2 years', '7 years', '8 years',\n",
       "       nan], dtype=object)"
      ]
     },
     "execution_count": 50,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['emp_length'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:54.630350Z",
     "start_time": "2019-01-26T14:22:54.153626Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "def encode_emp_length(x):\n",
    "    if type(x)==float: #Check for nan ,as only Nans are float while other values are str\n",
    "        return x\n",
    "    if x=='10+ years':\n",
    "        return 10\n",
    "    if x=='< 1 year':\n",
    "        return 0\n",
    "    else:\n",
    "        return int(x.split(' ')[0])\n",
    "    \n",
    "data['emp_length']=data['emp_length'].apply(encode_emp_length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:54.640340Z",
     "start_time": "2019-01-26T14:22:54.637346Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# for col in columns_with_nulls:\n",
    "#     plt.figure(figsize=(8,6))\n",
    "#     sns.distplot(data[data[col].notnull()][col])\n",
    "#     plt.title('Histogram of {}'.format(col))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Looking at the histograms, columns can be imputed by median value.\n",
    "I found that 'total_rev_hi_lim' is highly correlated with 'revol bal'.\n",
    "So, I had two methods to impute 'total_rev_hi_lim'.\n",
    "1. Impute by median value\n",
    "2. Fit a polynomial between 'total_rev_hi_lim' and 'revol_bal'\n",
    "\n",
    "As fitting a polynomial model didn't performed much better, I decided to go with method 1.\n",
    "Code for method 2 is commented below."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:55.115584Z",
     "start_time": "2019-01-26T14:22:54.643315Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "emp_length shows max correlation with total_acc : 0.14\n",
      "revol_util shows max correlation with int_rate : 0.34\n",
      "collections_12_mths_ex_med shows max correlation with issue_year : 0.06\n",
      "tot_coll_amt shows max correlation with revol_util : 0.01\n",
      "tot_cur_bal shows max correlation with annual_inc : 0.44\n",
      "total_rev_hi_lim shows max correlation with revol_bal : 0.83\n"
     ]
    }
   ],
   "source": [
    "corr_matrix=data.corr()\n",
    "for col in columns_with_nulls:\n",
    "    tmp=corr_matrix[col].apply(abs).sort_values(ascending=False)\n",
    "    print(col, 'shows max correlation with',tmp.index[1],':',round(tmp[1],2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:55.121569Z",
     "start_time": "2019-01-26T14:22:55.117593Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# Fit a linear mode between 'total_rev_hi_lim' and 'revol_bal'\n",
    "\n",
    "# from sklearn.metrics import mean_squared_error\n",
    "# from scipy import polyfit,poly1d\n",
    "\n",
    "# tmp1=data[data['total_rev_hi_lim'].notnull()][['revol_bal','total_rev_hi_lim']].copy()\n",
    "# p=polyfit(tmp1['revol_bal'],tmp1['total_rev_hi_lim'],deg=2)\n",
    "# poly=poly1d(p)\n",
    "# tmp1['total_rev_hi_lim_pred']=tmp1['revol_bal'].apply(poly)\n",
    "# print('RMSE:',np.sqrt(mean_squared_error(tmp1['total_rev_hi_lim'],tmp1['total_rev_hi_lim_pred'])))\n",
    "# data['temp']=data['revol_bal'].apply(poly)\n",
    "\n",
    "# i=data[data['total_rev_hi_lim'].isnull()].index\n",
    "# data.loc[i,'total_rev_hi_lim']=data.loc[i,'temp']\n",
    "# data.drop('temp',axis=1,inplace=True)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:00:43.827355Z",
     "start_time": "2019-01-26T14:00:43.823365Z"
    },
    "code_folding": []
   },
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:55.258950Z",
     "start_time": "2019-01-26T14:22:55.123562Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#Impute remaining columns with median\n",
    "\n",
    "def impute_by_median(data,col):\n",
    "    \"\"\"\n",
    "    Impute column by median value\n",
    "    Params:\n",
    "        data: Dataframe\n",
    "        col: str, Name of column\n",
    "    \"\"\"\n",
    "    med=data[col].median()\n",
    "    data[col]=data[col].apply(lambda x: med if np.isnan(x) else x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:57.007699Z",
     "start_time": "2019-01-26T14:22:55.260901Z"
    }
   },
   "outputs": [],
   "source": [
    "for col in columns_with_nulls:\n",
    "    impute_by_median(data,col)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:57.012261Z",
     "start_time": "2019-01-26T14:22:57.009273Z"
    }
   },
   "outputs": [],
   "source": [
    "# data.to_csv('DC2.csv',index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning 3: Miscellaneous \n",
    "I went thought each and every column to see whether it is useful or not and did necessasry processing.\n",
    "1. **'term'**: has only two values. I converted it to binary variable later.\n",
    "2. **'grade'**: has 7 values. I created dummy variables for it.\n",
    "3. **'home_ownership'**: Changed OHTER, NONE, ANY to RENT as these data points are too low.\n",
    "4. **'verification_status'**: has 3 unique values. Created dummy variables.\n",
    "5. **'pymnt_plan'**: only 6 rows have 'y' value. Dropped this column\n",
    "6. **'purpose'**: has 14 unique values. Created dummy variables for it.\n",
    "7. **'addr_state'**:\n",
    "\n",
    "    There are total 51 states. So creating 50 dummy variables is not a wise choice. So, I did detailed analysis for 'addr_state'\n",
    "    1. Some states having very less loans. So I considered states having more than 1000 loan applications.\n",
    "    2. I calculated default rate per state and default rate per year per state.\n",
    "    3. FL,LA,NV,IN : Higher default rate esp. after 2011\n",
    "    4. CA,NY : avg. default rate, but number of loans are high CA, NY\n",
    "    5. IL,GA,TX : Generally have lower default rate\n",
    "\n",
    "   I decided to create dummy variables for these states: 'FL', 'NY', 'LA', 'NV', 'IN', 'CA', 'IL', 'GA', 'TX'\n",
    "\n",
    "8. **'initial_list_status'**: don't seem to affect on 'loan_sttaus'. I dropped this column.\n",
    "9. **'application_type'**: Almost all the values fall in one category. I dropped this column\n",
    "10. **'earliest_cr_line'**: Changed it to the number of days before the loan issued i.e. number of days between 'issue_date' and 'earliest_cr_line'\n",
    "11. Divide **'annual_inc', 'tot_cur_bal', 'total_rev_hi_lim', 'revol_bal'** by **'loan_amnt'**.\n",
    "12. **'collections_12_mths_ex_med'**:  is very skewed with more than 90% values as 0.0. Also, default rate is similar for all values. I dropped this column as it is non-conclusive."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:57.133726Z",
     "start_time": "2019-01-26T14:22:57.017221Z"
    }
   },
   "outputs": [],
   "source": [
    "# data=pd.read_csv('DC2.csv')\n",
    "# data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:57.285167Z",
     "start_time": "2019-01-26T14:22:57.135704Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# data.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 60,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:57.423274Z",
     "start_time": "2019-01-26T14:22:57.287179Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array([' 36 months', ' 60 months'], dtype=object)"
      ]
     },
     "execution_count": 60,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['term'].unique()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:57.570294Z",
     "start_time": "2019-01-26T14:22:57.425270Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "B    78267\n",
       "C    69019\n",
       "D    43708\n",
       "A    42835\n",
       "E    21239\n",
       "F     8507\n",
       "G     2206\n",
       "Name: grade, dtype: int64"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data['grade'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:57.708523Z",
     "start_time": "2019-01-26T14:22:57.578266Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "MORTGAGE    130361\n",
       "RENT        111928\n",
       "OWN          23307\n",
       "OTHER          141\n",
       "NONE            43\n",
       "ANY              1\n",
       "Name: home_ownership, dtype: int64"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#home_ownership\n",
    "data['home_ownership'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:58.603129Z",
     "start_time": "2019-01-26T14:22:57.723483Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2850f69c898>"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x='home_ownership',hue='loan_status',data=data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:58.669034Z",
     "start_time": "2019-01-26T14:22:58.606121Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#Change OHTER, NONE, ANY to RENT as these data points are too low.\n",
    "data['home_ownership']=data['home_ownership'].apply(lambda x: 'RENT' if x in ['OTHER','NONE','ANY'] else x)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:58.757581Z",
     "start_time": "2019-01-26T14:22:58.670035Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Verified           97720\n",
       "Not Verified       88863\n",
       "Source Verified    79198\n",
       "Name: verification_status, dtype: int64"
      ]
     },
     "execution_count": 65,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#verification_status\n",
    "data['verification_status'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:59.382356Z",
     "start_time": "2019-01-26T14:22:58.759547Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2850f684a90>"
      ]
     },
     "execution_count": 66,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.countplot(x='verification_status',hue='loan_status',data=data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:59.479061Z",
     "start_time": "2019-01-26T14:22:59.384343Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "n    265775\n",
      "y         6\n",
      "Name: pymnt_plan, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "#pymnt_plan\n",
    "print(data['pymnt_plan'].value_counts())\n",
    "data.drop('pymnt_plan',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 68,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:59.581968Z",
     "start_time": "2019-01-26T14:22:59.481055Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Unique values in purpose column: 14\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "debt_consolidation    156572\n",
       "credit_card            52405\n",
       "home_improvement       15638\n",
       "other                  14937\n",
       "major_purchase          6486\n",
       "small_business          4955\n",
       "car                     3726\n",
       "medical                 2994\n",
       "moving                  2129\n",
       "wedding                 1963\n",
       "house                   1720\n",
       "vacation                1655\n",
       "educational              325\n",
       "renewable_energy         276\n",
       "Name: purpose, dtype: int64"
      ]
     },
     "execution_count": 68,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#purpose\n",
    "print(\"Unique values in purpose column:\",data['purpose'].nunique())\n",
    "data['purpose'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 69,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:59.774018Z",
     "start_time": "2019-01-26T14:22:59.584960Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>purpose_Default_Rate</th>\n",
       "      <th>purpose_Count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>car</th>\n",
       "      <td>0.141707</td>\n",
       "      <td>3726</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>credit_card</th>\n",
       "      <td>0.193779</td>\n",
       "      <td>52405</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>debt_consolidation</th>\n",
       "      <td>0.228700</td>\n",
       "      <td>156572</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>educational</th>\n",
       "      <td>0.172308</td>\n",
       "      <td>325</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>home_improvement</th>\n",
       "      <td>0.190434</td>\n",
       "      <td>15638</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>house</th>\n",
       "      <td>0.205814</td>\n",
       "      <td>1720</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>major_purchase</th>\n",
       "      <td>0.168825</td>\n",
       "      <td>6486</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>medical</th>\n",
       "      <td>0.236807</td>\n",
       "      <td>2994</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>moving</th>\n",
       "      <td>0.247064</td>\n",
       "      <td>2129</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>other</th>\n",
       "      <td>0.240744</td>\n",
       "      <td>14937</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>renewable_energy</th>\n",
       "      <td>0.228261</td>\n",
       "      <td>276</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>small_business</th>\n",
       "      <td>0.318870</td>\n",
       "      <td>4955</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>vacation</th>\n",
       "      <td>0.203625</td>\n",
       "      <td>1655</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>wedding</th>\n",
       "      <td>0.139073</td>\n",
       "      <td>1963</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    purpose_Default_Rate  purpose_Count\n",
       "car                             0.141707           3726\n",
       "credit_card                     0.193779          52405\n",
       "debt_consolidation              0.228700         156572\n",
       "educational                     0.172308            325\n",
       "home_improvement                0.190434          15638\n",
       "house                           0.205814           1720\n",
       "major_purchase                  0.168825           6486\n",
       "medical                         0.236807           2994\n",
       "moving                          0.247064           2129\n",
       "other                           0.240744          14937\n",
       "renewable_energy                0.228261            276\n",
       "small_business                  0.318870           4955\n",
       "vacation                        0.203625           1655\n",
       "wedding                         0.139073           1963"
      ]
     },
     "execution_count": 69,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Get the defalut rate for each laon purpose.\n",
    "get_default_rate_by_col('purpose')\n",
    "#Loans for small_business are highly probable to default."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:22:59.828846Z",
     "start_time": "2019-01-26T14:22:59.775997Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of unique values in addr_state column: 51\n"
     ]
    }
   ],
   "source": [
    "#addr_state\n",
    "print(\"Number of unique values in addr_state column:\",data['addr_state'].nunique())\n",
    "# data['addr_state'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:00.009996Z",
     "start_time": "2019-01-26T14:22:59.836832Z"
    }
   },
   "outputs": [],
   "source": [
    "default_rate_by_state=get_default_rate_by_col('addr_state')['addr_state_Default_Rate'].sort_values()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 72,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:00.021938Z",
     "start_time": "2019-01-26T14:23:00.012964Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "#Consider states with significant loan applications\n",
    "# default_rate_by_state[data['addr_state'].value_counts()>1000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:00.405467Z",
     "start_time": "2019-01-26T14:23:00.023959Z"
    }
   },
   "outputs": [],
   "source": [
    "#Check default rate per year per state\n",
    "tmp1=data.groupby(by=['issue_year','addr_state']).count()['loan_amnt']\n",
    "tmp2=data[data['loan_status']=='Default'].groupby(by=['issue_year','addr_state']).count()['loan_amnt']\n",
    "default_rate_by_year_state=pd.DataFrame(tmp2/tmp1)\n",
    "default_rate_by_year_state=default_rate_by_year_state.reset_index()\n",
    "default_rate_by_year_state=default_rate_by_year_state.sort_values(by=['issue_year','loan_amnt'],ascending=[True,False])\n",
    "default_rate_by_year_state=pd.DataFrame(default_rate_by_year_state)\n",
    "\n",
    "#Add count of rows for each state to the dataframe\n",
    "tmp=pd.DataFrame(data['addr_state'].value_counts())\n",
    "tmp.columns=['count']\n",
    "tmp.reset_index(inplace=True)\n",
    "tmp.columns=['addr_state','count']\n",
    "default_rate_by_year_state=pd.merge(default_rate_by_year_state,tmp,how='left',on='addr_state')\n",
    "pd.DataFrame(default_rate_by_year_state).to_csv('default_rate_by_yar_state.csv',index=False)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 74,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:01.269703Z",
     "start_time": "2019-01-26T14:23:00.406985Z"
    }
   },
   "outputs": [],
   "source": [
    "#From above dataframe\n",
    "states_to_keep=['FL','NY','LA','NV','IN','CA','IL','GA','TX']\n",
    "\n",
    "#Create dummy variables for these states\n",
    "for i in states_to_keep:\n",
    "    data['state'+i]=data['addr_state'].apply(lambda x: 1 if x==i else 0)\n",
    "data.drop('addr_state',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 75,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:01.849135Z",
     "start_time": "2019-01-26T14:23:01.271671Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "f    189062\n",
      "w     76719\n",
      "Name: initial_list_status, dtype: int64\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2850f78f400>"
      ]
     },
     "execution_count": 75,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#initial_list_status\n",
    "print(data['initial_list_status'].value_counts())\n",
    "sns.countplot(x='initial_list_status',hue='loan_status',data=data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:02.157302Z",
     "start_time": "2019-01-26T14:23:01.851123Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   initial_list_status_Default_Rate  initial_list_status_Count\n",
      "f                          0.211127                     189062\n",
      "w                          0.236473                      76719\n"
     ]
    }
   ],
   "source": [
    "print(get_default_rate_by_col('initial_list_status'))\n",
    "\n",
    "data.drop('initial_list_status',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:02.372727Z",
     "start_time": "2019-01-26T14:23:02.173312Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "INDIVIDUAL    265778\n",
      "JOINT              3\n",
      "Name: application_type, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "#application_type\n",
    "print(data['application_type'].value_counts())\n",
    "data.drop('application_type',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:07.387409Z",
     "start_time": "2019-01-26T14:23:02.381709Z"
    }
   },
   "outputs": [],
   "source": [
    "#earliest_cr_line\n",
    "\n",
    "#Change 'issue_d' and 'earliest_cr_line ' to datetime\n",
    "from datetime import datetime as dt\n",
    "data['issue_d']=data['issue_d'].apply(lambda x:dt.strptime('01-'+x,'%d-%b-%Y').date())\n",
    "data['earliest_cr_line']=data['earliest_cr_line'].apply(lambda x:dt.strptime('01-'+x,'%d-%b-%Y').date())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:11.288911Z",
     "start_time": "2019-01-26T14:23:07.390303Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "data['earliest_cr_line']=data['issue_d']-data['earliest_cr_line']\n",
    "data['earliest_cr_line']=data['earliest_cr_line'].apply(lambda x:int(str(x).split(' ')[0]))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 80,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:11.311839Z",
     "start_time": "2019-01-26T14:23:11.290897Z"
    }
   },
   "outputs": [],
   "source": [
    "#Divide 'annual_inc','tot_cur_bal','total_rev_hi_lim','revol_bal' by 'loan_amnt'\n",
    "tmp=['annual_inc','tot_cur_bal','total_rev_hi_lim','revol_bal']\n",
    "for col in tmp:\n",
    "    data[col]=data[col]/data['loan_amnt']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:11.881530Z",
     "start_time": "2019-01-26T14:23:11.313834Z"
    }
   },
   "outputs": [],
   "source": [
    "#Create dummy variables for categorical columns\n",
    "cat_cols=['term','grade','home_ownership','verification_status','purpose']\n",
    "for col in cat_cols:\n",
    "    data=data.join(pd.get_dummies(data[col],drop_first=True))\n",
    "data.drop(cat_cols,axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:11.887485Z",
     "start_time": "2019-01-26T14:23:11.883493Z"
    }
   },
   "outputs": [],
   "source": [
    "# collections_12_mths_ex_med\n",
    "# get_default_rate_by_col('collections_12_mths_ex_med')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 83,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:12.094444Z",
     "start_time": "2019-01-26T14:23:11.890477Z"
    }
   },
   "outputs": [],
   "source": [
    "data.drop('collections_12_mths_ex_med',axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 84,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:12.167819Z",
     "start_time": "2019-01-26T14:23:12.096438Z"
    }
   },
   "outputs": [],
   "source": [
    "# data.to_csv('DC3.csv',index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning 4: Handling Outliers \n",
    "In this section I have checked for the outliers. Outliers may force our machine learning model to overfit. Handling outliers features helps to decrease the effect of outliers on machine learning model.\n",
    "\n",
    "An easy way to detect outliers is to plot a box plot.\n",
    "\n",
    "There are many ways to handle the outliers like:\n",
    "1. Delete the rows with outliers\n",
    "2. Cap the features with outliers i.e. set some max value for that feature and cap the outliers at that value.\n",
    "3. Apply some transformations (mostly log or boxcox transformation) to feature and convert it to a near normal distribution.\n",
    "\n",
    "Instead of applying a single strategy to all features, I analysed them and then applied suitable strategy. If feature had extreme outliers, I considered deleting or capping those rows otherwise I applied transformations.\n",
    "\n",
    "Handling outliers itself can be another major task, but I have not much focused on it for this project."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:12.261449Z",
     "start_time": "2019-01-26T14:23:12.169691Z"
    }
   },
   "outputs": [],
   "source": [
    "# data=pd.read_csv('DC3.csv')\n",
    "# data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As an example I will explain how did I handle outliers for columns : \n",
    "\n",
    "**1. int_rate**\n",
    "\n",
    "#get_default_rate_by_col('int_rate')\n",
    "\n",
    "As interest rate increases, default rate is also increasing. So, outliers in this case can be helpful to predict defaulter.\n",
    "\n",
    "For interest rate > 25 around 50% loans are default.\n",
    "\n",
    "So, I've transformed this feature."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 86,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:12.476058Z",
     "start_time": "2019-01-26T14:23:12.263444Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2850f7c1c18>"
      ]
     },
     "execution_count": 86,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "#int_rate\n",
    "fig, ax = plt.subplots()\n",
    "fig.set_size_inches(12,6)\n",
    "sns.boxplot(x='int_rate',data=data,ax=ax)\n",
    "\n",
    "#The points outside the whiskers of boxplot represents outliers."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 87,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:12.490025Z",
     "start_time": "2019-01-26T14:23:12.479026Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Default       1116\n",
       "Fully Paid    1112\n",
       "Name: loan_status, dtype: int64"
      ]
     },
     "execution_count": 87,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "data[data['int_rate']>25]['loan_status'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 88,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:12.829629Z",
     "start_time": "2019-01-26T14:23:12.492989Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x28542e02b00>"
      ]
     },
     "execution_count": 88,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(data['int_rate'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 89,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:13.417080Z",
     "start_time": "2019-01-26T14:23:12.831620Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2850f679e10>"
      ]
     },
     "execution_count": 89,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(data['int_rate'].apply(lambda x: np.log1p(x)**2))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 90,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:13.797063Z",
     "start_time": "2019-01-26T14:23:13.419050Z"
    }
   },
   "outputs": [],
   "source": [
    "data['int_rate']=data['int_rate'].apply(lambda x: np.log1p(x)**2)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**2. dti**\n",
    "\n",
    "There is only one value greate than 40 which is 57.14. It is an extreme outlier, so I removed it. There is no need to apply any distribution."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 91,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:13.933673Z",
     "start_time": "2019-01-26T14:23:13.799057Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x28510794208>"
      ]
     },
     "execution_count": 91,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig, ax = plt.subplots()\n",
    "fig.set_size_inches(12,6)\n",
    "sns.boxplot(x='dti',data=data,ax=ax)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 92,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:14.083298Z",
     "start_time": "2019-01-26T14:23:13.936679Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "data=data[data['dti']<40].reset_index(drop=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 93,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:14.355578Z",
     "start_time": "2019-01-26T14:23:14.085268Z"
    }
   },
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x285108ffe48>"
      ]
     },
     "execution_count": 93,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "sns.distplot(data['dti'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this way I have handled outliers in the data. Some bullet points are: \n",
    "\n",
    "1. Instances with higher delinquency are too low to get any conclusion.\n",
    "2. As number of inquiries increases, we can see increase in default rate. But, higher 'inq_last_6mths' values have  too low instances. Same is the case for 'open_acc'\n",
    "3. Values with high 'revol_bal' have less defaulters. But values are much spreaded after 40.\n",
    "\n",
    "I can't mention all the steps here. I have added the final code of outlier removal in 'lending_club.data_cleaning.py' file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 94,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:17.390519Z",
     "start_time": "2019-01-26T14:23:14.358536Z"
    }
   },
   "outputs": [],
   "source": [
    "from lending_club.data_cleaning import Transform_Skewed_Features\n",
    "transformer=Transform_Skewed_Features()\n",
    "data=transformer.fit_transform(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 95,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:17.398407Z",
     "start_time": "2019-01-26T14:23:17.393425Z"
    }
   },
   "outputs": [],
   "source": [
    "# data.to_csv('DC4.csv',index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Data Cleaning 5: Multicolinearity\n",
    "\n",
    "I have also checked for multi-colinearity in data. Multicoliearity can be checked from variance Inflation Factor (vif). Basically, it detects highly correlated features from the data. Presence of highly correlated features in data will overfit the machine learning model. So, it is necessary to remove such features from data.\n",
    "\n",
    "But, removing all such features may not always help and even in worst case it can make the model worse. So, it is always good to train model in both cases or try changing thrshold of vif. I have not much focused on this part. \n",
    "\n",
    "I have already written code for this in my another project [California House Price Prediction](https://github.com/sonarsushant/California-House-Price-Prediction). I have added this code to 'lending_club.data_analysis.py' file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 96,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:17.517724Z",
     "start_time": "2019-01-26T14:23:17.401422Z"
    }
   },
   "outputs": [],
   "source": [
    "# data=pd.read_csv('DC4.csv')\n",
    "# data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 97,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:17.722685Z",
     "start_time": "2019-01-26T14:23:17.519603Z"
    }
   },
   "outputs": [],
   "source": [
    "y=data['loan_status'].apply(lambda x: 1 if x=='Default' else 0)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 98,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:17.798705Z",
     "start_time": "2019-01-26T14:23:17.724550Z"
    },
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "data.drop(['loan_status','issue_d','issue_year'],axis=1,inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 99,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:17.884899Z",
     "start_time": "2019-01-26T14:23:17.800700Z"
    }
   },
   "outputs": [],
   "source": [
    "from lending_club.data_cleaning import remove_by_vif"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 100,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:18.009905Z",
     "start_time": "2019-01-26T14:23:17.886871Z"
    },
    "code_folding": [],
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# remove_by_vif(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 101,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2019-01-26T14:23:18.109743Z",
     "start_time": "2019-01-26T14:23:18.017883Z"
    }
   },
   "outputs": [],
   "source": [
    "columns_removed_by_vif=['loan_amnt','int_rate','annual_inc','total_acc','revol_bal',\n",
    "                        'debt_consolidation','total_rev_hi_lim','tot_cur_bal','open_acc'\n",
    "                        'revol_util','earliest_cr_line',]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Next Steps:\n",
    "1. Prepare data for machine learning algorithm\n",
    "    1. Create a pipeline to do all the data cleaning I did in this notebook\n",
    "    2. Split data into train and test set\n",
    "    3. Applying SMOTE technique. (I have explained this in next notebook)\n",
    "    4. Standard scaling of data\n",
    "    \n",
    "    \n",
    "2. Training machine learning algorithms"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}