05_pos_balances.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5.0  Analysis of POS Balance\n",
    "\n",
    "> In this notebook I analyze the `pos cash balance` dataset which is described below:\n",
    "> - Monthly balance snapshots of previous point of sales (POS) and cash loans that the applicant had with Home Credit\n",
    "\n",
    "#### Analysis Summary:\n",
    "> After analyzing the `pos cash balance` dataset I created a vector for each loan id with following columns:\n",
    ">> - __COMPLETED_COUNT__: The number of completed accounts by bin size of 5( i.e. 1 means 0 to 5 accounts).\n",
    ">> - __OVERDUE_ACTIVE_COUNT__: Flag to indicate if a client has an overdue account.\n",
    "\n",
    "> I also created an aggregate table of four functions ('minimum', 'maximum', 'mean','standard deviation') `SK_ID_CURR` for each of the availble `float64` fields.  Each field would have a new column for each of the functions.  This created a total of 20 features.  By taking this approach I figure there will be many fields available during model selection that can be removed if there feature importance is zero.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "#import helper ultility module\n",
    "%load_ext autoreload\n",
    "%aimport utils\n",
    "%autoreload 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "pd.options.display.float_format = '{:,.2f}'.format\n",
    "pd.set_option('display.max_rows', 100)\n",
    "\n",
    "import numpy as np\n",
    "\n",
    "%matplotlib inline"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.1 Import Datasets And Prepare ABT\n",
    "> - Import the train and test datasets combined using a utility helper \n",
    "> - Create Activity Base Table (ABT) to group the fields by type."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset = utils.get_pos_cash_balances()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <th>CNT_INSTALMENT</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE</th>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <th>SK_DPD</th>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1803195</td>\n",
       "      <td>182943</td>\n",
       "      <td>-31.00</td>\n",
       "      <td>48.00</td>\n",
       "      <td>45.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1715348</td>\n",
       "      <td>367990</td>\n",
       "      <td>-33.00</td>\n",
       "      <td>36.00</td>\n",
       "      <td>35.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1784872</td>\n",
       "      <td>397406</td>\n",
       "      <td>-32.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>9.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1903291</td>\n",
       "      <td>269225</td>\n",
       "      <td>-35.00</td>\n",
       "      <td>48.00</td>\n",
       "      <td>42.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>2341044</td>\n",
       "      <td>334279</td>\n",
       "      <td>-35.00</td>\n",
       "      <td>36.00</td>\n",
       "      <td>35.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_PREV  SK_ID_CURR  MONTHS_BALANCE  CNT_INSTALMENT  \\\n",
       "0     1803195      182943          -31.00           48.00   \n",
       "1     1715348      367990          -33.00           36.00   \n",
       "2     1784872      397406          -32.00           12.00   \n",
       "3     1903291      269225          -35.00           48.00   \n",
       "4     2341044      334279          -35.00           36.00   \n",
       "\n",
       "   CNT_INSTALMENT_FUTURE NAME_CONTRACT_STATUS  SK_DPD  SK_DPD_DEF  \n",
       "0                  45.00               Active    0.00        0.00  \n",
       "1                  35.00               Active    0.00        0.00  \n",
       "2                   9.00               Active    0.00        0.00  \n",
       "3                  42.00               Active    0.00        0.00  \n",
       "4                  35.00               Active    0.00        0.00  "
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "abt = utils.ABT(dataset)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total float64 types: 5\n"
     ]
    },
    {
     "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>num</th>\n",
       "      <th>count</th>\n",
       "      <th>na_count</th>\n",
       "      <th>na_%</th>\n",
       "      <th>mean</th>\n",
       "      <th>std</th>\n",
       "      <th>min</th>\n",
       "      <th>25%</th>\n",
       "      <th>50%</th>\n",
       "      <th>75%</th>\n",
       "      <th>max</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <td>1</td>\n",
       "      <td>10,001,358.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>-35.01</td>\n",
       "      <td>26.07</td>\n",
       "      <td>-96.00</td>\n",
       "      <td>-54.00</td>\n",
       "      <td>-28.00</td>\n",
       "      <td>-13.00</td>\n",
       "      <td>-1.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT</th>\n",
       "      <td>2</td>\n",
       "      <td>9,975,287.00</td>\n",
       "      <td>26,071.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>17.09</td>\n",
       "      <td>12.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>10.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>92.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>CNT_INSTALMENT_FUTURE</th>\n",
       "      <td>3</td>\n",
       "      <td>9,975,271.00</td>\n",
       "      <td>26,087.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>10.48</td>\n",
       "      <td>11.11</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>14.00</td>\n",
       "      <td>85.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD</th>\n",
       "      <td>4</td>\n",
       "      <td>10,001,358.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>11.61</td>\n",
       "      <td>132.71</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>4,231.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "      <td>5</td>\n",
       "      <td>10,001,358.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.65</td>\n",
       "      <td>32.76</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3,595.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       num         count  na_count  na_%   mean    std    min  \\\n",
       "MONTHS_BALANCE           1 10,001,358.00      0.00  0.00 -35.01  26.07 -96.00   \n",
       "CNT_INSTALMENT           2  9,975,287.00 26,071.00  0.00  17.09  12.00   1.00   \n",
       "CNT_INSTALMENT_FUTURE    3  9,975,271.00 26,087.00  0.00  10.48  11.11   0.00   \n",
       "SK_DPD                   4 10,001,358.00      0.00  0.00  11.61 132.71   0.00   \n",
       "SK_DPD_DEF               5 10,001,358.00      0.00  0.00   0.65  32.76   0.00   \n",
       "\n",
       "                         25%    50%    75%      max  \n",
       "MONTHS_BALANCE        -54.00 -28.00 -13.00    -1.00  \n",
       "CNT_INSTALMENT         10.00  12.00  24.00    92.00  \n",
       "CNT_INSTALMENT_FUTURE   3.00   7.00  14.00    85.00  \n",
       "SK_DPD                  0.00   0.00   0.00 4,231.00  \n",
       "SK_DPD_DEF              0.00   0.00   0.00 3,595.00  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "float_table = abt.float_table\n",
    "print(\"Total float64 types: {}\".format(len(float_table)))\n",
    "float_table "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1) ['MONTHS_BALANCE']: Month of balance relative to application date (-1 means the information to the freshest monthly snapshot, 0 means the information at application - often it will be the same as -1 as many banks are not updating the information to Credit Bureau regularly )\n",
      "----------------------------------------\n",
      "2) ['CNT_INSTALMENT']: Term of previous credit (can change over time)\n",
      "----------------------------------------\n",
      "3) ['CNT_INSTALMENT_FUTURE']: Installments left to pay on the previous credit\n",
      "----------------------------------------\n",
      "4) ['SK_DPD']: DPD (days past due) during the month of previous credit\n",
      "----------------------------------------\n",
      "5) ['SK_DPD_DEF']: DPD during the month with tolerance (debts with low loan amounts are ignored) of the previous credit\n",
      "----------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# print description of each field in the integer ABT\n",
    "field_table = utils.FieldsNames()\n",
    "field_table.print('POS_CASH_balance.csv',float_table.index.values.tolist()) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total string types: 1\n"
     ]
    },
    {
     "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>num</th>\n",
       "      <th>count</th>\n",
       "      <th>na_count</th>\n",
       "      <th>na_%</th>\n",
       "      <th>unique</th>\n",
       "      <th>top</th>\n",
       "      <th>top_count</th>\n",
       "      <th>top_%</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>column</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <td>1</td>\n",
       "      <td>10001358</td>\n",
       "      <td>0</td>\n",
       "      <td>0.00</td>\n",
       "      <td>9</td>\n",
       "      <td>Active</td>\n",
       "      <td>9151119</td>\n",
       "      <td>0.91</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                      num     count  na_count  na_%  unique     top  \\\n",
       "column                                                                \n",
       "NAME_CONTRACT_STATUS    1  10001358         0  0.00       9  Active   \n",
       "\n",
       "                      top_count  top_%  \n",
       "column                                  \n",
       "NAME_CONTRACT_STATUS    9151119   0.91  "
      ]
     },
     "execution_count": 37,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "other_table = abt.other_table\n",
    "print(\"Total string types: {}\".format(len(other_table)))\n",
    "other_table "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "1) ['NAME_CONTRACT_STATUS']: Contract status during the month\n",
      "----------------------------------------\n"
     ]
    }
   ],
   "source": [
    "# print description of each field in the string ABT\n",
    "field_table = utils.FieldsNames()\n",
    "field_table.print('POS_CASH_balance.csv',other_table.index.values.tolist()) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x504 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "utils.plot_hbar(other_table.index,dataset)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.2 Filter Dataset\n",
    "> - Get the most recent record by `SK_ID_CURR` and `SK_ID_PREV`.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_filter = dataset.sort_values(by=['MONTHS_BALANCE']).groupby(['SK_ID_CURR', 'SK_ID_PREV']).last().copy().reset_index() "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <th>CNT_INSTALMENT</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE</th>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <th>SK_DPD</th>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>1369693</td>\n",
       "      <td>-53.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>Completed</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100001</td>\n",
       "      <td>1851984</td>\n",
       "      <td>-93.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>Completed</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100002</td>\n",
       "      <td>1038818</td>\n",
       "      <td>-1.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100003</td>\n",
       "      <td>1810518</td>\n",
       "      <td>-18.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>Completed</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100003</td>\n",
       "      <td>2396755</td>\n",
       "      <td>-66.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  SK_ID_PREV  MONTHS_BALANCE  CNT_INSTALMENT  \\\n",
       "0      100001     1369693          -53.00            4.00   \n",
       "1      100001     1851984          -93.00            4.00   \n",
       "2      100002     1038818           -1.00           24.00   \n",
       "3      100003     1810518          -18.00            7.00   \n",
       "4      100003     2396755          -66.00           12.00   \n",
       "\n",
       "   CNT_INSTALMENT_FUTURE NAME_CONTRACT_STATUS  SK_DPD  SK_DPD_DEF  \n",
       "0                   0.00            Completed    0.00        0.00  \n",
       "1                   0.00            Completed    0.00        0.00  \n",
       "2                   6.00               Active    0.00        0.00  \n",
       "3                   0.00            Completed    0.00        0.00  \n",
       "4                   1.00               Active    0.00        0.00  "
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_filter.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.3 Count Completed Installments\n",
    "> - Count the number of completed installments loans by id."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [],
   "source": [
    "def completed(df):\n",
    "    name_flag = 'COMPLETED_FLAG'\n",
    "    name = 'COMPLETED_COUNT'\n",
    "\n",
    "    df[name_flag] = df['NAME_CONTRACT_STATUS'].apply(lambda x: 1 if x =='Completed'  else 0)\n",
    "    \n",
    "    df[name] = df.groupby('SK_ID_CURR')[name_flag].transform('sum')\n",
    "    \n",
    "    #lets increase the groups by 1 to make the base group 1\n",
    "    df[name] = df[name].apply(lambda x:x+1)\n",
    "    \n",
    "    #drop column\n",
    "    df.drop([name_flag],axis=1, inplace=True)\n",
    "\n",
    "#apply function\n",
    "completed(dataset_filter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "SK_ID_CURR                 int64\n",
       "SK_ID_PREV                 int64\n",
       "MONTHS_BALANCE           float64\n",
       "CNT_INSTALMENT           float64\n",
       "CNT_INSTALMENT_FUTURE    float64\n",
       "NAME_CONTRACT_STATUS      object\n",
       "SK_DPD                   float64\n",
       "SK_DPD_DEF               float64\n",
       "COMPLETED_COUNT            int64\n",
       "dtype: object"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_filter.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 44,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "COMPLETED_COUNT\n",
       "1      43931\n",
       "2     182962\n",
       "3     204546\n",
       "4     168290\n",
       "5     121485\n",
       "6      81498\n",
       "7      51484\n",
       "8      31915\n",
       "9      20122\n",
       "10     11855\n",
       "11      7118\n",
       "12      4094\n",
       "13      2875\n",
       "14      1671\n",
       "15       945\n",
       "16       572\n",
       "17       369\n",
       "18       264\n",
       "19       177\n",
       "20        19\n",
       "21        62\n",
       "22        45\n",
       "25        26\n",
       "dtype: int64"
      ]
     },
     "execution_count": 44,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_filter.groupby('COMPLETED_COUNT').size()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.4 Count Overdue Active Counts\n",
    "> - Count the number of overdue active installments loans by id."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "def over_due_active(df):\n",
    "    ''' Count the number of overdue active accounts'''\n",
    "    \n",
    "    name_flag = 'OVERDUE_FLAG'\n",
    "    name = 'OVERDUE_ACTIVE_COUNT'\n",
    "    \n",
    "    def fun(x):\n",
    "        over = x['SK_DPD_DEF']\n",
    "        future_cnt = x['CNT_INSTALMENT_FUTURE']\n",
    "                       \n",
    "        if over > 0 and future_cnt > 0:\n",
    "            return 1\n",
    "        else:\n",
    "            return 0\n",
    "\n",
    "    df[name_flag] = df.apply(fun,axis=1)\n",
    "    \n",
    "    df[name] = df.groupby('SK_ID_CURR')[name_flag].transform('sum')\n",
    "    \n",
    "    #lets increase the groups by 1 to make the base group 1\n",
    "    df[name] = df[name].apply(lambda x:x+1)\n",
    "    \n",
    "    #drop column\n",
    "    df.drop([name_flag],axis=1, inplace=True)\n",
    "\n",
    "#apply function\n",
    "over_due_active(dataset_filter)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>SK_ID_PREV</th>\n",
       "      <th>MONTHS_BALANCE</th>\n",
       "      <th>CNT_INSTALMENT</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE</th>\n",
       "      <th>NAME_CONTRACT_STATUS</th>\n",
       "      <th>SK_DPD</th>\n",
       "      <th>SK_DPD_DEF</th>\n",
       "      <th>COMPLETED_COUNT</th>\n",
       "      <th>OVERDUE_ACTIVE_COUNT</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>1369693</td>\n",
       "      <td>-53.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>Completed</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100001</td>\n",
       "      <td>1851984</td>\n",
       "      <td>-93.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>Completed</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100002</td>\n",
       "      <td>1038818</td>\n",
       "      <td>-1.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100003</td>\n",
       "      <td>1810518</td>\n",
       "      <td>-18.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>Completed</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100003</td>\n",
       "      <td>2396755</td>\n",
       "      <td>-66.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>1.00</td>\n",
       "      <td>Active</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  SK_ID_PREV  MONTHS_BALANCE  CNT_INSTALMENT  \\\n",
       "0      100001     1369693          -53.00            4.00   \n",
       "1      100001     1851984          -93.00            4.00   \n",
       "2      100002     1038818           -1.00           24.00   \n",
       "3      100003     1810518          -18.00            7.00   \n",
       "4      100003     2396755          -66.00           12.00   \n",
       "\n",
       "   CNT_INSTALMENT_FUTURE NAME_CONTRACT_STATUS  SK_DPD  SK_DPD_DEF  \\\n",
       "0                   0.00            Completed    0.00        0.00   \n",
       "1                   0.00            Completed    0.00        0.00   \n",
       "2                   6.00               Active    0.00        0.00   \n",
       "3                   0.00            Completed    0.00        0.00   \n",
       "4                   1.00               Active    0.00        0.00   \n",
       "\n",
       "   COMPLETED_COUNT  OVERDUE_ACTIVE_COUNT  \n",
       "0                3                     1  \n",
       "1                3                     1  \n",
       "2                1                     1  \n",
       "3                3                     1  \n",
       "4                3                     1  "
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_filter.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OVERDUE_ACTIVE_COUNT\n",
       "1    930577\n",
       "2      5172\n",
       "3       450\n",
       "4       110\n",
       "5        16\n",
       "dtype: int64"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_filter.groupby('OVERDUE_ACTIVE_COUNT').size()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.5 Create Aggregate Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "#let get minimum, maximum, average and standard deviation\n",
    "fun_agg = ['min', 'max', 'mean','std']\n",
    "\n",
    "agg_fun_df = dataset.groupby('SK_ID_CURR')[float_table.index.values].agg(fun_agg)\n",
    "\n",
    "#rename columns\n",
    "agg_fun_df.columns = [\"{}_{}\".format(c[0],c[1].upper()) for c in agg_fun_df.columns.values.tolist()] \n",
    "\n",
    "#let fill nan values\n",
    "agg_fun_df.fillna(0,inplace=True)\n",
    "agg_fun_df.reset_index(inplace=True)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>MONTHS_BALANCE_MIN</th>\n",
       "      <th>MONTHS_BALANCE_MAX</th>\n",
       "      <th>MONTHS_BALANCE_MEAN</th>\n",
       "      <th>MONTHS_BALANCE_STD</th>\n",
       "      <th>CNT_INSTALMENT_MIN</th>\n",
       "      <th>CNT_INSTALMENT_MAX</th>\n",
       "      <th>CNT_INSTALMENT_MEAN</th>\n",
       "      <th>CNT_INSTALMENT_STD</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE_MIN</th>\n",
       "      <th>...</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE_MEAN</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE_STD</th>\n",
       "      <th>SK_DPD_MIN</th>\n",
       "      <th>SK_DPD_MAX</th>\n",
       "      <th>SK_DPD_MEAN</th>\n",
       "      <th>SK_DPD_STD</th>\n",
       "      <th>SK_DPD_DEF_MIN</th>\n",
       "      <th>SK_DPD_DEF_MAX</th>\n",
       "      <th>SK_DPD_DEF_MEAN</th>\n",
       "      <th>SK_DPD_DEF_STD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>-96.00</td>\n",
       "      <td>-53.00</td>\n",
       "      <td>-72.56</td>\n",
       "      <td>20.86</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>1.44</td>\n",
       "      <td>1.42</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0.78</td>\n",
       "      <td>2.33</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0.78</td>\n",
       "      <td>2.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>-19.00</td>\n",
       "      <td>-1.00</td>\n",
       "      <td>-10.00</td>\n",
       "      <td>5.63</td>\n",
       "      <td>24.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>6.00</td>\n",
       "      <td>...</td>\n",
       "      <td>15.00</td>\n",
       "      <td>5.63</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>-77.00</td>\n",
       "      <td>-18.00</td>\n",
       "      <td>-43.79</td>\n",
       "      <td>24.64</td>\n",
       "      <td>6.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>10.11</td>\n",
       "      <td>2.81</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>5.79</td>\n",
       "      <td>3.84</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>-27.00</td>\n",
       "      <td>-24.00</td>\n",
       "      <td>-25.50</td>\n",
       "      <td>1.29</td>\n",
       "      <td>3.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>3.75</td>\n",
       "      <td>0.50</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>2.25</td>\n",
       "      <td>1.71</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>-25.00</td>\n",
       "      <td>-15.00</td>\n",
       "      <td>-20.00</td>\n",
       "      <td>3.32</td>\n",
       "      <td>9.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>11.70</td>\n",
       "      <td>0.95</td>\n",
       "      <td>0.00</td>\n",
       "      <td>...</td>\n",
       "      <td>7.20</td>\n",
       "      <td>3.61</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  MONTHS_BALANCE_MIN  MONTHS_BALANCE_MAX  MONTHS_BALANCE_MEAN  \\\n",
       "0      100001              -96.00              -53.00               -72.56   \n",
       "1      100002              -19.00               -1.00               -10.00   \n",
       "2      100003              -77.00              -18.00               -43.79   \n",
       "3      100004              -27.00              -24.00               -25.50   \n",
       "4      100005              -25.00              -15.00               -20.00   \n",
       "\n",
       "   MONTHS_BALANCE_STD  CNT_INSTALMENT_MIN  CNT_INSTALMENT_MAX  \\\n",
       "0               20.86                4.00                4.00   \n",
       "1                5.63               24.00               24.00   \n",
       "2               24.64                6.00               12.00   \n",
       "3                1.29                3.00                4.00   \n",
       "4                3.32                9.00               12.00   \n",
       "\n",
       "   CNT_INSTALMENT_MEAN  CNT_INSTALMENT_STD  CNT_INSTALMENT_FUTURE_MIN  ...  \\\n",
       "0                 4.00                0.00                       0.00  ...   \n",
       "1                24.00                0.00                       6.00  ...   \n",
       "2                10.11                2.81                       0.00  ...   \n",
       "3                 3.75                0.50                       0.00  ...   \n",
       "4                11.70                0.95                       0.00  ...   \n",
       "\n",
       "   CNT_INSTALMENT_FUTURE_MEAN  CNT_INSTALMENT_FUTURE_STD  SK_DPD_MIN  \\\n",
       "0                        1.44                       1.42        0.00   \n",
       "1                       15.00                       5.63        0.00   \n",
       "2                        5.79                       3.84        0.00   \n",
       "3                        2.25                       1.71        0.00   \n",
       "4                        7.20                       3.61        0.00   \n",
       "\n",
       "   SK_DPD_MAX  SK_DPD_MEAN  SK_DPD_STD  SK_DPD_DEF_MIN  SK_DPD_DEF_MAX  \\\n",
       "0        7.00         0.78        2.33            0.00            7.00   \n",
       "1        0.00         0.00        0.00            0.00            0.00   \n",
       "2        0.00         0.00        0.00            0.00            0.00   \n",
       "3        0.00         0.00        0.00            0.00            0.00   \n",
       "4        0.00         0.00        0.00            0.00            0.00   \n",
       "\n",
       "   SK_DPD_DEF_MEAN  SK_DPD_DEF_STD  \n",
       "0             0.78            2.33  \n",
       "1             0.00            0.00  \n",
       "2             0.00            0.00  \n",
       "3             0.00            0.00  \n",
       "4             0.00            0.00  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "agg_fun_df.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.6 Select Final Features and Merge Datasets\n",
    "> - Select the `COMPLETED_COUNT`, `OVERDUE_ACTIVE_COUNT` and all from aggregate table as final features."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "features= ['COMPLETED_COUNT','OVERDUE_ACTIVE_COUNT']\n",
    "dataset_final = dataset_filter.groupby('SK_ID_CURR').first()[features].reset_index()\n",
    "\n",
    "#set max values\n",
    "utils.applyMax(['OVERDUE_ACTIVE_COUNT'],dataset_final,percent=99.9)\n",
    "utils.applyMax(['COMPLETED_COUNT'],dataset_final)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {},
   "outputs": [],
   "source": [
    "dataset_final = dataset_final.merge(agg_fun_df,\n",
    "                                      how='left', \n",
    "                                      on='SK_ID_CURR',\n",
    "                                      validate='one_to_one')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>SK_ID_CURR</th>\n",
       "      <th>COMPLETED_COUNT</th>\n",
       "      <th>OVERDUE_ACTIVE_COUNT</th>\n",
       "      <th>MONTHS_BALANCE_MIN</th>\n",
       "      <th>MONTHS_BALANCE_MAX</th>\n",
       "      <th>MONTHS_BALANCE_MEAN</th>\n",
       "      <th>MONTHS_BALANCE_STD</th>\n",
       "      <th>CNT_INSTALMENT_MIN</th>\n",
       "      <th>CNT_INSTALMENT_MAX</th>\n",
       "      <th>CNT_INSTALMENT_MEAN</th>\n",
       "      <th>...</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE_MEAN</th>\n",
       "      <th>CNT_INSTALMENT_FUTURE_STD</th>\n",
       "      <th>SK_DPD_MIN</th>\n",
       "      <th>SK_DPD_MAX</th>\n",
       "      <th>SK_DPD_MEAN</th>\n",
       "      <th>SK_DPD_STD</th>\n",
       "      <th>SK_DPD_DEF_MIN</th>\n",
       "      <th>SK_DPD_DEF_MAX</th>\n",
       "      <th>SK_DPD_DEF_MEAN</th>\n",
       "      <th>SK_DPD_DEF_STD</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>100001</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>-96.00</td>\n",
       "      <td>-53.00</td>\n",
       "      <td>-72.56</td>\n",
       "      <td>20.86</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>...</td>\n",
       "      <td>1.44</td>\n",
       "      <td>1.42</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0.78</td>\n",
       "      <td>2.33</td>\n",
       "      <td>0.00</td>\n",
       "      <td>7.00</td>\n",
       "      <td>0.78</td>\n",
       "      <td>2.33</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>100002</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>-19.00</td>\n",
       "      <td>-1.00</td>\n",
       "      <td>-10.00</td>\n",
       "      <td>5.63</td>\n",
       "      <td>24.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>24.00</td>\n",
       "      <td>...</td>\n",
       "      <td>15.00</td>\n",
       "      <td>5.63</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>100003</td>\n",
       "      <td>3</td>\n",
       "      <td>1</td>\n",
       "      <td>-77.00</td>\n",
       "      <td>-18.00</td>\n",
       "      <td>-43.79</td>\n",
       "      <td>24.64</td>\n",
       "      <td>6.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>10.11</td>\n",
       "      <td>...</td>\n",
       "      <td>5.79</td>\n",
       "      <td>3.84</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>100004</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>-27.00</td>\n",
       "      <td>-24.00</td>\n",
       "      <td>-25.50</td>\n",
       "      <td>1.29</td>\n",
       "      <td>3.00</td>\n",
       "      <td>4.00</td>\n",
       "      <td>3.75</td>\n",
       "      <td>...</td>\n",
       "      <td>2.25</td>\n",
       "      <td>1.71</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>100005</td>\n",
       "      <td>2</td>\n",
       "      <td>1</td>\n",
       "      <td>-25.00</td>\n",
       "      <td>-15.00</td>\n",
       "      <td>-20.00</td>\n",
       "      <td>3.32</td>\n",
       "      <td>9.00</td>\n",
       "      <td>12.00</td>\n",
       "      <td>11.70</td>\n",
       "      <td>...</td>\n",
       "      <td>7.20</td>\n",
       "      <td>3.61</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "      <td>0.00</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 23 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   SK_ID_CURR  COMPLETED_COUNT  OVERDUE_ACTIVE_COUNT  MONTHS_BALANCE_MIN  \\\n",
       "0      100001                3                     1              -96.00   \n",
       "1      100002                1                     1              -19.00   \n",
       "2      100003                3                     1              -77.00   \n",
       "3      100004                2                     1              -27.00   \n",
       "4      100005                2                     1              -25.00   \n",
       "\n",
       "   MONTHS_BALANCE_MAX  MONTHS_BALANCE_MEAN  MONTHS_BALANCE_STD  \\\n",
       "0              -53.00               -72.56               20.86   \n",
       "1               -1.00               -10.00                5.63   \n",
       "2              -18.00               -43.79               24.64   \n",
       "3              -24.00               -25.50                1.29   \n",
       "4              -15.00               -20.00                3.32   \n",
       "\n",
       "   CNT_INSTALMENT_MIN  CNT_INSTALMENT_MAX  CNT_INSTALMENT_MEAN  ...  \\\n",
       "0                4.00                4.00                 4.00  ...   \n",
       "1               24.00               24.00                24.00  ...   \n",
       "2                6.00               12.00                10.11  ...   \n",
       "3                3.00                4.00                 3.75  ...   \n",
       "4                9.00               12.00                11.70  ...   \n",
       "\n",
       "   CNT_INSTALMENT_FUTURE_MEAN  CNT_INSTALMENT_FUTURE_STD  SK_DPD_MIN  \\\n",
       "0                        1.44                       1.42        0.00   \n",
       "1                       15.00                       5.63        0.00   \n",
       "2                        5.79                       3.84        0.00   \n",
       "3                        2.25                       1.71        0.00   \n",
       "4                        7.20                       3.61        0.00   \n",
       "\n",
       "   SK_DPD_MAX  SK_DPD_MEAN  SK_DPD_STD  SK_DPD_DEF_MIN  SK_DPD_DEF_MAX  \\\n",
       "0        7.00         0.78        2.33            0.00            7.00   \n",
       "1        0.00         0.00        0.00            0.00            0.00   \n",
       "2        0.00         0.00        0.00            0.00            0.00   \n",
       "3        0.00         0.00        0.00            0.00            0.00   \n",
       "4        0.00         0.00        0.00            0.00            0.00   \n",
       "\n",
       "   SK_DPD_DEF_MEAN  SK_DPD_DEF_STD  \n",
       "0             0.78            2.33  \n",
       "1             0.00            0.00  \n",
       "2             0.00            0.00  \n",
       "3             0.00            0.00  \n",
       "4             0.00            0.00  \n",
       "\n",
       "[5 rows x 23 columns]"
      ]
     },
     "execution_count": 52,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_final.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "OVERDUE_ACTIVE_COUNT\n",
       "1    335764\n",
       "2      1488\n",
       "dtype: int64"
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_final.groupby('OVERDUE_ACTIVE_COUNT').size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "COMPLETED_COUNT\n",
       "1     37041\n",
       "2    122019\n",
       "3     78201\n",
       "4     45111\n",
       "5     25111\n",
       "6     13738\n",
       "7      7333\n",
       "8      3933\n",
       "9      4765\n",
       "dtype: int64"
      ]
     },
     "execution_count": 54,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "dataset_final.groupby('COMPLETED_COUNT').size()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x504 with 4 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "utils.plot_bar(features,dataset_final)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 56,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "['COMPLETED_COUNT',\n",
       " 'OVERDUE_ACTIVE_COUNT',\n",
       " 'MONTHS_BALANCE_MIN',\n",
       " 'MONTHS_BALANCE_MAX',\n",
       " 'MONTHS_BALANCE_MEAN',\n",
       " 'MONTHS_BALANCE_STD',\n",
       " 'CNT_INSTALMENT_MIN',\n",
       " 'CNT_INSTALMENT_MAX',\n",
       " 'CNT_INSTALMENT_MEAN',\n",
       " 'CNT_INSTALMENT_STD',\n",
       " 'CNT_INSTALMENT_FUTURE_MIN',\n",
       " 'CNT_INSTALMENT_FUTURE_MAX',\n",
       " 'CNT_INSTALMENT_FUTURE_MEAN',\n",
       " 'CNT_INSTALMENT_FUTURE_STD',\n",
       " 'SK_DPD_MIN',\n",
       " 'SK_DPD_MAX',\n",
       " 'SK_DPD_MEAN',\n",
       " 'SK_DPD_STD',\n",
       " 'SK_DPD_DEF_MIN',\n",
       " 'SK_DPD_DEF_MAX',\n",
       " 'SK_DPD_DEF_MEAN',\n",
       " 'SK_DPD_DEF_STD']"
      ]
     },
     "execution_count": 56,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pos_cols_int = ['POS_BAL_COMPLETED_COUNT', 'POS_BAL_OVERDUE_ACTIVE_COUNT']\n",
    "\n",
    "[f for f in dataset_final.columns if f != 'SK_ID_CURR' and f not in pos_cols_int]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 5.7 Save Final Dataset To File"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 57,
   "metadata": {},
   "outputs": [],
   "source": [
    "# rename columns  append all fields with 'POS_BAL_' except id field\n",
    "\n",
    "col_name = []\n",
    "for c in dataset_final.columns:\n",
    "    if c =='SK_ID_CURR':\n",
    "        col_name.append(c)\n",
    "    else:\n",
    "        col_name.append(\"POS_BAL_\"+c)\n",
    "\n",
    "dataset_final.columns = col_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "File was saved.\n"
     ]
    }
   ],
   "source": [
    "dataset_final.to_csv('data_pre_processed/1_5_pos_balances.csv',index=False)\n",
    "print(\"File was saved.\")"
   ]
  }
 ],
 "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}