ANZ - Virtual Experience Programs - Forage - Task 2 Predictive Analytics.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "9fae30e0",
   "metadata": {},
   "source": [
    "# Load Required Libraries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "b7354e67",
   "metadata": {},
   "outputs": [],
   "source": [
    "# import libraries\n",
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "\n",
    "from sklearn.model_selection import train_test_split\n",
    "from sklearn.linear_model import LinearRegression\n",
    "from sklearn.tree import DecisionTreeRegressor\n",
    "from sklearn import metrics"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "29874b49",
   "metadata": {},
   "source": [
    "# Load Dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "fda7d9d7",
   "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>status</th>\n",
       "      <th>card_present_flag</th>\n",
       "      <th>txn_description</th>\n",
       "      <th>merchant_id</th>\n",
       "      <th>balance</th>\n",
       "      <th>date</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>merchant_suburb</th>\n",
       "      <th>merchant_state</th>\n",
       "      <th>transaction_date_time</th>\n",
       "      <th>amount</th>\n",
       "      <th>customer_id</th>\n",
       "      <th>movement</th>\n",
       "      <th>customer_longitude</th>\n",
       "      <th>customer_latitude</th>\n",
       "      <th>merchant_longitude</th>\n",
       "      <th>merchant_latitude</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>authorized</td>\n",
       "      <td>1</td>\n",
       "      <td>POS</td>\n",
       "      <td>81c48296-73be-44a7-befa-d053f48ce7cd</td>\n",
       "      <td>35.39</td>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>F</td>\n",
       "      <td>26</td>\n",
       "      <td>Ashmore</td>\n",
       "      <td>QLD</td>\n",
       "      <td>2018-08-01 01:01:15+00:00</td>\n",
       "      <td>16.25</td>\n",
       "      <td>CUS-2487424745</td>\n",
       "      <td>debit</td>\n",
       "      <td>153.41</td>\n",
       "      <td>-27.95</td>\n",
       "      <td>153.38</td>\n",
       "      <td>-27.99</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>authorized</td>\n",
       "      <td>0</td>\n",
       "      <td>SALES-POS</td>\n",
       "      <td>830a451c-316e-4a6a-bf25-e37caedca49e</td>\n",
       "      <td>21.20</td>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>F</td>\n",
       "      <td>26</td>\n",
       "      <td>Sydney</td>\n",
       "      <td>NSW</td>\n",
       "      <td>2018-08-01 01:13:45+00:00</td>\n",
       "      <td>14.19</td>\n",
       "      <td>CUS-2487424745</td>\n",
       "      <td>debit</td>\n",
       "      <td>153.41</td>\n",
       "      <td>-27.95</td>\n",
       "      <td>151.21</td>\n",
       "      <td>-33.87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>authorized</td>\n",
       "      <td>1</td>\n",
       "      <td>POS</td>\n",
       "      <td>835c231d-8cdf-4e96-859d-e9d571760cf0</td>\n",
       "      <td>5.71</td>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>M</td>\n",
       "      <td>38</td>\n",
       "      <td>Sydney</td>\n",
       "      <td>NSW</td>\n",
       "      <td>2018-08-01 01:26:15+00:00</td>\n",
       "      <td>6.42</td>\n",
       "      <td>CUS-2142601169</td>\n",
       "      <td>debit</td>\n",
       "      <td>151.23</td>\n",
       "      <td>-33.94</td>\n",
       "      <td>151.21</td>\n",
       "      <td>-33.87</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>authorized</td>\n",
       "      <td>1</td>\n",
       "      <td>SALES-POS</td>\n",
       "      <td>48514682-c78a-4a88-b0da-2d6302e64673</td>\n",
       "      <td>2117.22</td>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>F</td>\n",
       "      <td>40</td>\n",
       "      <td>Buderim</td>\n",
       "      <td>QLD</td>\n",
       "      <td>2018-08-01 01:38:45+00:00</td>\n",
       "      <td>40.90</td>\n",
       "      <td>CUS-1614226872</td>\n",
       "      <td>debit</td>\n",
       "      <td>153.10</td>\n",
       "      <td>-27.66</td>\n",
       "      <td>153.05</td>\n",
       "      <td>-26.68</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>authorized</td>\n",
       "      <td>1</td>\n",
       "      <td>SALES-POS</td>\n",
       "      <td>b4e02c10-0852-4273-b8fd-7b3395e32eb0</td>\n",
       "      <td>17.95</td>\n",
       "      <td>2018-08-01</td>\n",
       "      <td>F</td>\n",
       "      <td>26</td>\n",
       "      <td>Mermaid Beach</td>\n",
       "      <td>QLD</td>\n",
       "      <td>2018-08-01 01:51:15+00:00</td>\n",
       "      <td>3.25</td>\n",
       "      <td>CUS-2487424745</td>\n",
       "      <td>debit</td>\n",
       "      <td>153.41</td>\n",
       "      <td>-27.95</td>\n",
       "      <td>153.44</td>\n",
       "      <td>-28.06</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "       status  card_present_flag txn_description  \\\n",
       "0  authorized                  1             POS   \n",
       "1  authorized                  0       SALES-POS   \n",
       "2  authorized                  1             POS   \n",
       "3  authorized                  1       SALES-POS   \n",
       "4  authorized                  1       SALES-POS   \n",
       "\n",
       "                            merchant_id  balance        date gender  age  \\\n",
       "0  81c48296-73be-44a7-befa-d053f48ce7cd    35.39  2018-08-01      F   26   \n",
       "1  830a451c-316e-4a6a-bf25-e37caedca49e    21.20  2018-08-01      F   26   \n",
       "2  835c231d-8cdf-4e96-859d-e9d571760cf0     5.71  2018-08-01      M   38   \n",
       "3  48514682-c78a-4a88-b0da-2d6302e64673  2117.22  2018-08-01      F   40   \n",
       "4  b4e02c10-0852-4273-b8fd-7b3395e32eb0    17.95  2018-08-01      F   26   \n",
       "\n",
       "  merchant_suburb merchant_state      transaction_date_time  amount  \\\n",
       "0         Ashmore            QLD  2018-08-01 01:01:15+00:00   16.25   \n",
       "1          Sydney            NSW  2018-08-01 01:13:45+00:00   14.19   \n",
       "2          Sydney            NSW  2018-08-01 01:26:15+00:00    6.42   \n",
       "3         Buderim            QLD  2018-08-01 01:38:45+00:00   40.90   \n",
       "4   Mermaid Beach            QLD  2018-08-01 01:51:15+00:00    3.25   \n",
       "\n",
       "      customer_id movement  customer_longitude  customer_latitude  \\\n",
       "0  CUS-2487424745    debit              153.41             -27.95   \n",
       "1  CUS-2487424745    debit              153.41             -27.95   \n",
       "2  CUS-2142601169    debit              151.23             -33.94   \n",
       "3  CUS-1614226872    debit              153.10             -27.66   \n",
       "4  CUS-2487424745    debit              153.41             -27.95   \n",
       "\n",
       "   merchant_longitude  merchant_latitude  \n",
       "0              153.38             -27.99  \n",
       "1              151.21             -33.87  \n",
       "2              151.21             -33.87  \n",
       "3              153.05             -26.68  \n",
       "4              153.44             -28.06  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# import dataset saved from task 1\n",
    "anz_data = pd.read_csv('ANZ_data_clean.csv')\n",
    "anz_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0fa8aad8",
   "metadata": {},
   "source": [
    "# Calculate Annual Salaries"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "07ef7495",
   "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>customer_id</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>F</td>\n",
       "      <td>53</td>\n",
       "      <td>12616.11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>M</td>\n",
       "      <td>21</td>\n",
       "      <td>25050.55</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>M</td>\n",
       "      <td>28</td>\n",
       "      <td>11499.06</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>F</td>\n",
       "      <td>34</td>\n",
       "      <td>22248.07</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>F</td>\n",
       "      <td>34</td>\n",
       "      <td>27326.11</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id gender  age    amount\n",
       "0  CUS-1005756958      F   53  12616.11\n",
       "1  CUS-1117979751      M   21  25050.55\n",
       "2  CUS-1140341822      M   28  11499.06\n",
       "3  CUS-1147642491      F   34  22248.07\n",
       "4  CUS-1196156254      F   34  27326.11"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# compute total salary received in 3 months for each customers\n",
    "salary_data = anz_data[anz_data['txn_description'] == 'PAY/SALARY'].groupby(['customer_id', 'gender', 'age']).amount.sum().reset_index()\n",
    "salary_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "0466b523",
   "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>customer_id</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>annual_salary</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>F</td>\n",
       "      <td>53</td>\n",
       "      <td>50464.44</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>M</td>\n",
       "      <td>21</td>\n",
       "      <td>100202.20</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>M</td>\n",
       "      <td>28</td>\n",
       "      <td>45996.24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>F</td>\n",
       "      <td>34</td>\n",
       "      <td>88992.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>F</td>\n",
       "      <td>34</td>\n",
       "      <td>109304.44</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id gender  age  annual_salary\n",
       "0  CUS-1005756958      F   53       50464.44\n",
       "1  CUS-1117979751      M   21      100202.20\n",
       "2  CUS-1140341822      M   28       45996.24\n",
       "3  CUS-1147642491      F   34       88992.28\n",
       "4  CUS-1196156254      F   34      109304.44"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculating annual salary (multiply salary received in 3 months with 4)\n",
    "salary_data = salary_data.assign(amount=salary_data['amount'] * 4)\n",
    "\n",
    "salary_data = salary_data.rename(columns = {'amount' : 'annual_salary'})\n",
    "\n",
    "salary_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1d244dbe",
   "metadata": {},
   "source": [
    "# Feature Creation"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7a705e8a",
   "metadata": {},
   "source": [
    "From the existing data let us create some additional features to 'age' and 'gender' which could help us predict the annual salary of the customer.\n",
    "\n",
    "The below are the additional features created based on the existing variables:\n",
    "* Average Purchase Amount\n",
    "* Minimum Purchase Amount\n",
    "* Maximum Purchase Amount\n",
    "* Average Number of Purchases made in a month\n",
    "* Average Balance"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "a7ca3b81",
   "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>customer_id</th>\n",
       "      <th>avg_purchase_amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>37.726250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>76.458077</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>67.531385</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>51.128289</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>30.310491</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id  avg_purchase_amount\n",
       "0  CUS-1005756958            37.726250\n",
       "1  CUS-1117979751            76.458077\n",
       "2  CUS-1140341822            67.531385\n",
       "3  CUS-1147642491            51.128289\n",
       "4  CUS-1196156254            30.310491"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate average purchase amount for each customer\n",
    "avg_purchase_amount = anz_data[anz_data['txn_description'].isin(['POS', 'SALES-POS'])].groupby(['customer_id']).amount.mean().reset_index()\n",
    "avg_purchase_amount = avg_purchase_amount.rename(columns = {'amount' : 'avg_purchase_amount'})\n",
    "avg_purchase_amount.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "c80fa54d",
   "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>customer_id</th>\n",
       "      <th>min_purchase_amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>6.48</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>5.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>4.09</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>8.78</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>3.66</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id  min_purchase_amount\n",
       "0  CUS-1005756958                 6.48\n",
       "1  CUS-1117979751                 5.64\n",
       "2  CUS-1140341822                 4.09\n",
       "3  CUS-1147642491                 8.78\n",
       "4  CUS-1196156254                 3.66"
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate minimum purchase amount for each customer\n",
    "min_purchase_amount = anz_data[anz_data['txn_description'].isin(['POS', 'SALES-POS'])].groupby(['customer_id']).amount.min().reset_index()\n",
    "min_purchase_amount = min_purchase_amount.rename(columns = {'amount' : 'min_purchase_amount'})\n",
    "min_purchase_amount.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "9cb41a70",
   "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>customer_id</th>\n",
       "      <th>max_purchase_amount</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>226.82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>2885.85</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>1271.46</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>432.92</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>390.67</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id  max_purchase_amount\n",
       "0  CUS-1005756958               226.82\n",
       "1  CUS-1117979751              2885.85\n",
       "2  CUS-1140341822              1271.46\n",
       "3  CUS-1147642491               432.92\n",
       "4  CUS-1196156254               390.67"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate maximum purchase amount for each customer\n",
    "max_purchase_amount = anz_data[anz_data['txn_description'].isin(['POS', 'SALES-POS'])].groupby(['customer_id']).amount.max().reset_index()\n",
    "max_purchase_amount = max_purchase_amount.rename(columns = {'amount' : 'max_purchase_amount'})\n",
    "max_purchase_amount.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "96cc69e7",
   "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>customer_id</th>\n",
       "      <th>avg_monthly_purchases</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>16</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>21</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>25</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>54</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id  avg_monthly_purchases\n",
       "0  CUS-1005756958                     16\n",
       "1  CUS-1117979751                     17\n",
       "2  CUS-1140341822                     21\n",
       "3  CUS-1147642491                     25\n",
       "4  CUS-1196156254                     54"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate average number of purchases made by customer each month\n",
    "num_of_purchases = anz_data[anz_data['txn_description'].isin(['POS', 'SALES-POS'])].groupby(['customer_id']).amount.count().reset_index()\n",
    "num_of_purchases = num_of_purchases.assign(amount= num_of_purchases['amount'] / 3)\n",
    "num_of_purchases['amount'] = num_of_purchases['amount'].astype(int)\n",
    "num_of_purchases = num_of_purchases.rename(columns = {'amount' : 'avg_monthly_purchases'})\n",
    "num_of_purchases.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "3b6a6787",
   "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>customer_id</th>\n",
       "      <th>avg_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>2275.852055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>9829.929000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>5699.212250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>9032.841186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>22272.433755</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id   avg_balance\n",
       "0  CUS-1005756958   2275.852055\n",
       "1  CUS-1117979751   9829.929000\n",
       "2  CUS-1140341822   5699.212250\n",
       "3  CUS-1147642491   9032.841186\n",
       "4  CUS-1196156254  22272.433755"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# calculate average balance for each customer\n",
    "avg_balance = anz_data.groupby(['customer_id']).balance.mean().reset_index()\n",
    "avg_balance = avg_balance.rename(columns = {'balance' : 'avg_balance'})\n",
    "avg_balance.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "273cc710",
   "metadata": {},
   "source": [
    "# Combine all features into one dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "afc0f78b",
   "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>customer_id</th>\n",
       "      <th>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>annual_salary</th>\n",
       "      <th>avg_purchase_amount</th>\n",
       "      <th>min_purchase_amount</th>\n",
       "      <th>max_purchase_amount</th>\n",
       "      <th>avg_monthly_purchases</th>\n",
       "      <th>avg_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>CUS-1005756958</td>\n",
       "      <td>F</td>\n",
       "      <td>53</td>\n",
       "      <td>50464.44</td>\n",
       "      <td>37.726250</td>\n",
       "      <td>6.48</td>\n",
       "      <td>226.82</td>\n",
       "      <td>16</td>\n",
       "      <td>2275.852055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CUS-1117979751</td>\n",
       "      <td>M</td>\n",
       "      <td>21</td>\n",
       "      <td>100202.20</td>\n",
       "      <td>76.458077</td>\n",
       "      <td>5.64</td>\n",
       "      <td>2885.85</td>\n",
       "      <td>17</td>\n",
       "      <td>9829.929000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>CUS-1140341822</td>\n",
       "      <td>M</td>\n",
       "      <td>28</td>\n",
       "      <td>45996.24</td>\n",
       "      <td>67.531385</td>\n",
       "      <td>4.09</td>\n",
       "      <td>1271.46</td>\n",
       "      <td>21</td>\n",
       "      <td>5699.212250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>CUS-1147642491</td>\n",
       "      <td>F</td>\n",
       "      <td>34</td>\n",
       "      <td>88992.28</td>\n",
       "      <td>51.128289</td>\n",
       "      <td>8.78</td>\n",
       "      <td>432.92</td>\n",
       "      <td>25</td>\n",
       "      <td>9032.841186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>CUS-1196156254</td>\n",
       "      <td>F</td>\n",
       "      <td>34</td>\n",
       "      <td>109304.44</td>\n",
       "      <td>30.310491</td>\n",
       "      <td>3.66</td>\n",
       "      <td>390.67</td>\n",
       "      <td>54</td>\n",
       "      <td>22272.433755</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      customer_id gender  age  annual_salary  avg_purchase_amount  \\\n",
       "0  CUS-1005756958      F   53       50464.44            37.726250   \n",
       "1  CUS-1117979751      M   21      100202.20            76.458077   \n",
       "2  CUS-1140341822      M   28       45996.24            67.531385   \n",
       "3  CUS-1147642491      F   34       88992.28            51.128289   \n",
       "4  CUS-1196156254      F   34      109304.44            30.310491   \n",
       "\n",
       "   min_purchase_amount  max_purchase_amount  avg_monthly_purchases  \\\n",
       "0                 6.48               226.82                     16   \n",
       "1                 5.64              2885.85                     17   \n",
       "2                 4.09              1271.46                     21   \n",
       "3                 8.78               432.92                     25   \n",
       "4                 3.66               390.67                     54   \n",
       "\n",
       "    avg_balance  \n",
       "0   2275.852055  \n",
       "1   9829.929000  \n",
       "2   5699.212250  \n",
       "3   9032.841186  \n",
       "4  22272.433755  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# merging salary_data and other dataframes created during Feature Creation\n",
    "final_data = pd.merge(salary_data, avg_purchase_amount, on='customer_id', how='inner')\n",
    "final_data = pd.merge(final_data, min_purchase_amount, on='customer_id', how='inner')\n",
    "final_data = pd.merge(final_data, max_purchase_amount, on='customer_id', how='inner')\n",
    "final_data = pd.merge(final_data, num_of_purchases, on='customer_id', how='inner')\n",
    "final_data = pd.merge(final_data, avg_balance, on='customer_id', how='inner')\n",
    "final_data.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "feacc7b0",
   "metadata": {},
   "source": [
    "# Explore Correlations"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "15024911",
   "metadata": {},
   "outputs": [],
   "source": [
    "# returns correlation between only relevant variables, i.e annual salary vs any other features\n",
    "def correlation(df):\n",
    "    correlation = df.corr()\n",
    "    correlation.drop(columns = ['annual_salary'], inplace = True)\n",
    "    correlation.drop(['age', 'avg_purchase_amount', 'min_purchase_amount', 'max_purchase_amount', 'avg_monthly_purchases', 'avg_balance'], inplace = True)\n",
    "    return correlation    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "a22a2edd",
   "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>age</th>\n",
       "      <th>avg_purchase_amount</th>\n",
       "      <th>min_purchase_amount</th>\n",
       "      <th>max_purchase_amount</th>\n",
       "      <th>avg_monthly_purchases</th>\n",
       "      <th>avg_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>annual_salary</th>\n",
       "      <td>-0.036504</td>\n",
       "      <td>-0.000346</td>\n",
       "      <td>-0.033626</td>\n",
       "      <td>0.011574</td>\n",
       "      <td>-0.058413</td>\n",
       "      <td>0.186986</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    age  avg_purchase_amount  min_purchase_amount  \\\n",
       "annual_salary -0.036504            -0.000346            -0.033626   \n",
       "\n",
       "               max_purchase_amount  avg_monthly_purchases  avg_balance  \n",
       "annual_salary             0.011574              -0.058413     0.186986  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# display correlation between annual salary and other features for all customers\n",
    "correlation(final_data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "1240b13e",
   "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>age</th>\n",
       "      <th>avg_purchase_amount</th>\n",
       "      <th>min_purchase_amount</th>\n",
       "      <th>max_purchase_amount</th>\n",
       "      <th>avg_monthly_purchases</th>\n",
       "      <th>avg_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>annual_salary</th>\n",
       "      <td>0.007593</td>\n",
       "      <td>0.057464</td>\n",
       "      <td>-0.053716</td>\n",
       "      <td>0.104556</td>\n",
       "      <td>-0.176983</td>\n",
       "      <td>0.077251</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    age  avg_purchase_amount  min_purchase_amount  \\\n",
       "annual_salary  0.007593             0.057464            -0.053716   \n",
       "\n",
       "               max_purchase_amount  avg_monthly_purchases  avg_balance  \n",
       "annual_salary             0.104556              -0.176983     0.077251  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# display correlation between annual salary and other features for all male customers\n",
    "correlation(final_data[final_data['gender'] == 'M'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "2e106236",
   "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>age</th>\n",
       "      <th>avg_purchase_amount</th>\n",
       "      <th>min_purchase_amount</th>\n",
       "      <th>max_purchase_amount</th>\n",
       "      <th>avg_monthly_purchases</th>\n",
       "      <th>avg_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>annual_salary</th>\n",
       "      <td>-0.107625</td>\n",
       "      <td>-0.15415</td>\n",
       "      <td>-0.041724</td>\n",
       "      <td>-0.08268</td>\n",
       "      <td>0.061253</td>\n",
       "      <td>0.360151</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                    age  avg_purchase_amount  min_purchase_amount  \\\n",
       "annual_salary -0.107625             -0.15415            -0.041724   \n",
       "\n",
       "               max_purchase_amount  avg_monthly_purchases  avg_balance  \n",
       "annual_salary             -0.08268               0.061253     0.360151  "
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# display correlation between annual salary and other features for all female customers\n",
    "correlation(final_data[final_data['gender'] == 'F'])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "2e3cbe82",
   "metadata": {},
   "source": [
    "annual salary shares very weak correlation with other attributes of the customer. The highest correlation among all is 0.187, i.e. between annual salary and customer's average balance. Although correlation is weak but it still suggests few interesting things:\n",
    "* Male customer's salary tends to increase with age unlike female customers.\n",
    "* Female tends spend less and save more as their salary increases unlike male customers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "98ea5dd6",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAZoAAAEGCAYAAABcolNbAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAABNX0lEQVR4nO3dd3xUVfr48c+T3nshhZAAAQxSBATsWLGXXQtWXFnZ5qrbdd3f7qrrVtft6te1FxR07b1gLygISu8tECCk93p+f5wbUmYCKTOZmeR5v17zYubk3jvnAplnTnuOGGNQSimlvCXI1xVQSik1uGmgUUop5VUaaJRSSnmVBhqllFJepYFGKaWUV4X4ugL+JCUlxeTm5vq6GkopFVCWLVu23xiT2t3PNdB0kJuby9KlS31dDaWUCigisv1gP9euM6WUUl6lgUYppZRXaaBRSinlVTpGo5RSPtTU1ERhYSH19fW+rsohRUREkJ2dTWhoaK/O00CjlFI+VFhYSGxsLLm5uYiIr6vTLWMMJSUlFBYWkpeX16tzNdCooa1sOxStgMZaSDsMhk2EIO1RVgOnvr7e74MMgIiQnJxMcXFxr8/VQKOGrtItsOAS2L/Bvg4KgSufg7zjfVsvNeT4e5Bp09d6evWrm4g8KCL7RGSVm5/9VESMiKR0KLtZRDaJyHoRmd2hfKqIrHR+9k9x7lZEwkVkoVO+RERyO5wzV0Q2Oo+53rxPFaAKl7YHGYDWZnjndmio8l2dlBqEvN1H8DBwetdCERkOnArs6FBWAMwBxjvn3C0iwc6P7wHmA/nOo+2a84AyY8xo4G/An5xrJQG/AWYA04HfiEiih+9NBbpqN10A5duhsW7g66LUALj66qt55plnBvx9vRpojDEfAKVufvQ34OdAx13XzgOeMsY0GGO2ApuA6SKSAcQZYz41dpe2R4HzO5zziPP8GeBkp7UzG3jLGFNqjCkD3sJNwFNDXNYRrmVT5kJMt5k0lBpSmpubPXKdAR/1FJFzgV3GmK+6/CgL2NnhdaFTluU871re6RxjTDNQASQf5FpKtcs8Ai56BOKyICQCjroOplwFAdJfrga322+/nXHjxnHqqady6aWXcuedd7J582ZOP/10pk6dynHHHce6desA21K5/vrrOfrooxk5cuSBVosxhuuuu46CggLOOuss9u3bd+D6y5Yt44QTTmDq1KnMnj2boqIiAGbNmsUvf/lLTjjhBP7xj3945F4GdDKAiEQBtwCnufuxmzJzkPK+ntO1TvOx3XLk5OS4O0QNVqGRMP58GHE0tDRCbAYEBR/yNKW8benSpfzvf/9j+fLlNDc3M2XKFKZOncr8+fO59957yc/PZ8mSJXz/+99n8eLFABQVFfHRRx+xbt06zj33XC688EKee+451q9fz8qVK9m7dy8FBQVcc801NDU18cMf/pAXXniB1NRUFi5cyC233MKDDz4IQHl5Oe+//77H7megZ52NAvKAr5zx/GzgSxGZjm11DO9wbDaw2ynPdlNOh3MKRSQEiMd21RUCs7qc8567Chlj7gPuA5g2bZrbYKQGuZg0X9dAqU4++ugjzjvvPCIjIwE455xzqK+v55NPPuGiiy46cFxDQ8OB5+effz5BQUEUFBSwd+9eAD744AMuvfRSgoODyczM5KSTTgJg/fr1rFq1ilNPPRWAlpYWMjIyDlzrkksu8ej9DGigMcasBA78VovINmCaMWa/iLwILBCRu4BM7KD/58aYFhGpEpGZwBLgKuBfziVeBOYCnwIXAouNMUZE3gB+32ECwGnAzd6/Q6WU6j87HN1Za2srCQkJrFixwu054eHhbs93NyXZGMP48eP59NNP3V4rOjq6lzU+OG9Pb34SGwTGikihiMzr7lhjzGpgEbAGeB34gTGmxfnx94D7sRMENgOvOeUPAMkisgn4MXCTc61S4HbgC+dxm1OmlFJ+79hjj+Wll16ivr6e6upqXnnlFaKiosjLy+Ppp58GbLD46quuQ92dHX/88Tz11FO0tLRQVFTEu+++C8DYsWMpLi4+EGiamppYvXq11+7Hqy0aY8ylh/h5bpfXdwB3uDluKXC4m/J64KKu5c7PHgQe7EV1lVLKLxx55JGce+65TJo0iREjRjBt2jTi4+N54okn+N73vsfvfvc7mpqamDNnDpMmTer2OhdccAGLFy9mwoQJjBkzhhNOOAGAsLAwnnnmGa6//noqKipobm7mxhtvZPz48V65H3HXRBuqpk2bZnTjM6XUQFq7di2HHXaYS3l1dTUxMTHU1tZy/PHHc9999zFlyhQf1LAzd/UVkWXGmGndnaMpaJTqq4pdUPgFlG6FYRMgexpEJvi6VmqQmD9/PmvWrKG+vp65c+f6RZDpKw00SvVFzX548Yew+Z32shN/Bcf+CIL110r134IFC3xdBY/RNLVK9cW+tZ2DDMAHf4ayrb6pj1J+TAONUn3R5CYfWksjNDe4lis1xGkbX/mnxlrYtRQ2L7YLKkeeaPeL8Rcp+RCZCHVl7WV5J0DiCN/VSSk/pYFG+adNb8OiK9tfR6fA1a9C6ljf1amjpDy7d817f4LdX8Jh58KM70J4rK9rppTf0UCj/E9tGbxza+eymv2wa5n/BBpwknI+ZPeviUzSSQAqYAUHBzNhwoQDr59//nlyc3M9dn39zVD+p7UJGqtdy92Ni/haaKR9KBXAIiMju01t4wk6GUD5n5g0OOqHncuCQyErcNcRKOUpzy/fxTF/XEzeTa9wzB8X8/zyXb6u0iFpi0b5p4kXQ1gUfH6f3S/muB/DsO5TbSg1FDy/fBc3P7uSuiabBnJXeR03P7sSgPOP6PuWW3V1dUyePBmAvLw8nnvuuX7XtSMNNMo/xaTBtGtgwkUQHAYh4Yc+R6lB7i9vrD8QZNrUNbXwlzfW9yvQeLvrTAON8m86i0upA3aXux+n7K7cX+gYjVJKBYjMBPcTT7or9xcaaJTf2lxczVur9/Dxpv3sr9IV90r9bPZYIkM7bzceGRrMz2b70bR/N7TrTPmlL3eUceX9S6hptP3Rx41O4c8XTiTDz7+5KeVNbeMwf3ljPbvL68hMiORns8f2a3wG7JYE3qSBRvmdmoZm/vzaugNBBuDDTfv5qrBcA40a8s4/IqvfgWWgadeZ8jvVDc2s3VPlUl5UUe+D2iil+ksDjfI7SdFhnDlhmEv5mHSdgaZUINJAo/xOaHAQ848fyayxqYAd7Pz12QVMyo73cc2UUn2hYzTq0JrqbeLIqCQICj708R6QlxLDfy6bwq7yOsJDgshJikJEBuS9lVKepYFGHdzu5fDen2HPCii4AI6cB8mjBuSto8NDtLtMqUFAu85U90q3wGMXwIZXoXI3fPYfeOOX0Fjj65oppTxIRLjyyvb9n5qbm0lNTeXss8/2yPU10KjuFW/ovIMkwIbXoWy7b+qjlPKK6OhoVq1aRV2dTWXz1ltvkZXluSnUGmhU90IjXMtCwiEkbODropSyvl4Efzscfptg//x6kUcue8YZZ/DKK68A8OSTT3LppZd65LqggUYdTFoBjDi2c9kJN0Finm/qo9RQ9/UieOl6qNgJGPvnS9d7JNjMmTOHp556ivr6er7++mtmzJjR//o6dDKA6l5MGlxwL+z83I7XZE6G7CMHbOaZUqqLd25z3Wm2qc6WT7y4X5eeOHEi27Zt48knn+TMM8/s17W68mqLRkQeFJF9IrKqQ9lfRGSdiHwtIs+JSEKHn90sIptEZL2IzO5QPlVEVjo/+6c481xFJFxEFjrlS0Qkt8M5c0Vko/OY6837HNQShsOEb8IJP4P8UyEywdc1UmroqijsXXkvnXvuufz0pz/1aLcZeL/r7GHg9C5lbwGHG2MmAhuAmwFEpACYA4x3zrlbRNq+Ot8DzAfynUfbNecBZcaY0cDfgD8510oCfgPMAKYDvxGRRC/cn1JKDZz47N6V99I111zDr3/9ayZMmOCR67XxaqAxxnwAlHYpe9MY0+y8/Axo+xs6D3jKGNNgjNkKbAKmi0gGEGeM+dQYY4BHgfM7nPOI8/wZ4GSntTMbeMsYU2qMKcMGt64BTynbJbj6BVjxFOz6ElqaD32OUr5y8q8htEti2dBIW+4B2dnZ3HDDDR65Vke+HqO5BljoPM/CBp42hU5Zk/O8a3nbOTsBjDHNIlIBJHcsd3NOJyIyH9taIicnpx+3ogJOyRY2rlnG59XplDUmMr1yH5OalhGe67lBUKU8qm0c5p3bbHdZfLYNMv0cn3G3TcCsWbOYNWtWv67bxmeBRkRuAZqBJ9qK3BxmDlLe13M6FxpzH3AfwLRp09weowanTftrufT9BPZXt2eKfuDcIE4eVgkRcT6smVIHMfHifgeWgeaT6c3O4PzZwOVOdxjYVsfwDodlA7ud8mw35Z3OEZEQIB7bVdfdtZQ64Mti2F/d2KnsziU1VNQ1+ahGSg1OAx5oROR04BfAucaY2g4/ehGY48wky8MO+n9ujCkCqkRkpjP+chXwQodz2maUXQgsdgLXG8BpIpLoTAI4zSlT6oCaFtcGfXl9C02hMT6ojRrK2r9v+7e+1tOrXWci8iQwC0gRkULsTLCbgXDgLWeW8mfGmO8aY1aLyCJgDbZL7QfGmLYtFr+HncEWCbzmPAAeAB4TkU3YlswcAGNMqYjcDnzhHHebMabTpASPaaiCoq+hbBvEZkDGRIhO8cpbKc+anJtKkGymtcPvzryjc0iJCfddpdSQExERQUlJCcnJyX6dodwYQ0lJCRERbjKGHIIESiQdCNOmTTNLly7t+QmtLfDZPfDmLe1lU66G027XPv4A0NzSypKtpfz97Q3sq2rgW0fncubEDNJie/+LpFRfNTU1UVhYSH29/+8gGxERQXZ2NqGhoZ3KRWSZMWZad+dpoOmg14GmeAPcewy0dO7nZ95bMHy6ZyunvKa2sZmm5lbiozSHm1J9cahA4+vpzYGtsdo1yADUVw58XVSfRYWFgMYYpbxGk2r2R0IOpI7rXBYRD8kjfVMfpZTyQxpo+iM6BS58AEadAhIEmVPg8qchSQONUkq10a6z/ko/HC55FGpKbGsmMt7XNVJKKb+igcYTwqLto4PWVkNZbSMx4SGEh2pafaXU0KWBxgu2l9TwxJLtvLiiiPGZcVx/cj6Thif4ulpKKeUTGmg8rK6xhT+8uo7XV+8BYE9lPZ9vLeWF645hZKquOFdKDT06GcDDCstrDwSZNlUNzWza55odVSmlhgJt0XhYWHAQEaFB1De1diqP7O84TdVeKNkIQSGQMgaikvp3PT+zs7SWHaW1xEWEMCo1hqhw/a+p1GChv80eNjwxih+dOoY/vLruQNnUEQmMzYjt+0WL18OiuVC81r7OPQHO+zckDo79c1bsLOOah5dSWmMXv84/Po/vzxpNgq7UV2pQ0EDjYUFBwpwjhzNuWCxf7awgJymKabmJfc+fZQysWNAeZAC2vQ9b3oWpc7s/L0BU1jVx64trDgQZgPs+2MoJY9I4ZrQmJ1VqMNBA4wXxkWGcMCaNE8ak9f9izQ2w5T3X8h2fDYpAU17XxPKd5S7lRRX+n2BQKdUzOhnA34VGwNgzXctHnjDwdfGCpOhQZuS5jjdlJ0a6OVopFYg00ASCiRdB7vHtrydcDHnHd398AIkJD+X/nV1AVoINLMFBwk9PG8P4DN1mQanBQrcJ6KDX2wQMpLoKKN1sZ50lj4SwwbUmZ19lPTtKa4mNCGFkSgyhIfodyGtKNsGeldDaCukFkHaYr2ukApxuEzBYRMZD1hRf18Jr0uIiSIvTDce8bu8aePRcqCm2r8PjYO6LkHmEb+ulBjX92qjUULL+1fYgA9BQCUsfsrMblfISDTRKDSX7N7qWFa+F1uaBr4saMjTQKDWUHHaOa9kRV0FwqGu5Uh6igUapoWTEMXDGnyEiwW5tceItMOY0X9dKDXI6GUCpoSQqEWZ8x7ZsTCvEZYGIr2ulBjkNNGrIWL+nki+2lVHX2My03CQmZicQHDREP2TjMn1dAzWE9DjQiMjhxphV3qyMUt6yrqiSS+77jIq6JsAuDH183gxm5CWxs6yWxuZWMhMjiQ7T715KeVpvfqvuFZEw4GFggTGm3Cs1UsoLPthYfCDIALS0Gt5Zu4fVuyv465sbqGtq4eRxafzq7MPISxlci2GV8rUeTwYwxhwLXA4MB5aKyAIROdVrNRsqSrfBprdh5xKoK/d1bQatspoml7Jh8ZH87pW11DW1APDOun08/PE2mltaXY5VSvVdr2adGWM2Ar8CfgGcAPxTRNaJyDfcHS8iD4rIPhFZ1aEsSUTeEpGNzp+JHX52s4hsEpH1IjK7Q/lUEVnp/OyfInb0UkTCRWShU75ERHI7nDPXeY+NIuKfaY53fQn/nQWPfxMeOA1e/and4Ex53AljUju9jg4LprrBde3IKyuLOm1ZoJTqvx4HGhGZKCJ/A9YCJwHnGGMOc57/rZvTHgZO71J2E/COMSYfeMd5jYgUAHOA8c45d4tI27aU9wDzgXzn0XbNeUCZMWa0U4c/OddKAn4DzACmA7/pGND8QkM1vP1bqCtrL1v5NBSt8FWNBrUjchK4f+40xmfGMSI5il+ddRgjU6JdjivIiCMmQsdplPKk3vxG/Rv4L/BLY0xdW6ExZreI/MrdCcaYDzq2MhznAbOc548A72FbSOcBTxljGoCtIrIJmC4i24A4Y8ynACLyKHA+8Jpzzm+daz0D/Ntp7cwG3jLGlDrnvIUNTk/24n69q74Cdi1zLS/fOfB18QNb91ezo6SW2MhQ8lNjiI307ALC8NBgTjksnRl5SbS0GhKiwthdXsfMkUl8tqUUgJjwEH506hiidEKAUh7Vo98op2Wx0xjzmLufd1fejXRjTJFzXpGItO0OlgV81uG4QqesyXnetbztnJ3OtZpFpAJI7lju5pxORGQ+trVETs4Abo0cnQL5p8HqZzuXp4weuDr4iaXbSrn6oS8OdGVdOn04P5s9lqTo8L5ftLkRyrbaHF6JuXZfHyA2oj2AZSZE8q9Lp7B+byV1DS2MSothZKpOBFDK03oUaIwxLSKSLCJhxhhvdWC7W9BgDlLe13M6FxpzH3Af2G0CDl1NDwkJhxN+4aRs/9qmADn+F77NottUDyUb7ZYEiTmQ4P3AW1HbyG9eXN1pvOTJz3dy1oQMjs1PPciZB1G5Bz66C5bebxclTrocTrwZ4rNdDk2NDSc1to/vo5Tqkd70EWwHPhaRF4GatkJjzF29fM+9IpLhtGYygH1OeSF2RlubbGC3U57tprzjOYUiEgLEA6VO+awu57zXy3p6X9o4uOoFKNsGoVGQPMp3OafqK+Gze+D9P9hWQFQyXPoUDJ/u1betrG9m9e5Kl/I9lQ19v+iWd+Hz/2t/veJxyJwM06/t+zWVUn3Wm1lnu4GXnXNiOzx660WgbRbYXOCFDuVznJlkedhB/8+dbrYqEZnpjL9c1eWctmtdCCw2die3N4DTRCTRmQRwmlPmf6KS7D4zaeN8m9hw7yp47/ft6eJrS+ClG6G27KCn9VdSdBjHjU5xKc9J6sdWzutecS1b+bTd6EspNeB63KIxxtza24uLyJPYlkWKiBRiZ4L9EVgkIvOAHcBFzvVXi8giYA3QDPzAGNPiXOp72BlskdhJAK855Q8AjzkTB0qxs9YwxpSKyO3AF85xt7VNDFDdqNjlWrZvNdSW2vxYXhIdHsIvzzqMHz75JZv21RAeEsTPTx/L+Mz4vl80exqse6lz2YijIUhzyCrlCz3eyllEUoGfY6cfH9gK0RhzkneqNvD8eitnb9v+KTzUZSb6sAlw1UteDTRtSqob2FVWR3RECLnJ0f3LQbZ/Izx1GezfYF/H58Dli3TLYqW8xJNbOT8BLATOBr6L7bIqPugZKnAMOxxO/g28+ztobYGYNDj77wMSZACSY8JJjunHLLOOUvLt2Ne+tXYyQOo4SBh+6POUUl7Rm0CTbIx5QERuMMa8D7wvIu97q2JqgIXHwlHXwZjZNhVO4gi3s7QCRlymZihWyk/0JtC0JYsqEpGzsJMDAviTyI81N0JFoZ0cMJDfxEPCIH38wL2fUmpI6E2g+Z2IxAM/Af4FxAE/8kqthrKyHfDhnXZKbmi07c6aeDFExPm6Zkop1Se9mXX2svO0AjjRO9VRrHgCvnzEPm+ohFd/Akl5MPpk39ZLKaX66JCBRkT+RTer6gGMMdd7tEZDWW0pfOUmHduOJRpolFIBqyctmiE639cHQqPsFNzy7Z3Lk/Ioqqhj075qQoKE/LRYUmI9NENLKaW87JCBxhjzyEBURGETPx7/U9j2ITQ6WX7SCtiQfCLz7v2UnWU2afbk4Qn8Y85kRiS7prlXXnIgD1wZJIyws/KUUj3S4zEaZ8HmL4ACBumCTb+QfSRcu9iuAQmJxGRMYtEHpQeCDMCKneV8uHG/BpqB0lAFn90L791hU/REJsKlT0LOUb6umVIBoTc5OZ7AbnqWB9wKbKM9xYvypNRxMP4CGHs6DZFpfLa1xOWQFTvLB75eQ9XeVXYha1sWjboyeOGHULPft/VSKkD0JtAkG2MeAJqMMe8bY64BZnqpXsoRERrMGYdnuJQf6yYRZaCqrGti8dq9/L/nV/HAh1vYtK/K11XqzF0euJKNdvKGUuqQdMGmF7S0Gjbtq2ZHaQ1J0WGMTY8lJqLvmZnPmZTJVzvKeXPtXkTgsuk5HDUq2YM19q2XvtrNLc+vOvD6vx9tYeH8o/yna9DdvjypBXbzOqXUIemCTS94b/0+vvv4MppabFfLtcflcf1J+X3enjgnKYq/zZnM9pIagoOE3JRowkOCPVlln9lTUc9f3lzfpayB1bsr+xdoGqqgZLPt7koe1b8Fr2nj4bQ74O3fQGuzzQN37j/tFg9KqUPSBZseVlRex03/W3kgyAD898OtnDZ+GEfm9v2DKTo8hIL+pM73Uy2trdQ3tbiUN7f0Y++Y8p3w5q9gzfP29ehT4My/QlJu364XHg3TvwOjT7K7jybkQLzbncGVUm70eIxGRP4sInEiEioi74jIfhG5wpuVC0SV9c0UV7vuDllc1Y8dIwexjPhI5h83slNZVFgw44b1owWy6e32INP2eu2Lfb8eQEgopBXAiKM0yCjVS72ZDHCaMaYSu01AITAG+JlXahXA0uPCGTus88ajIjAiKcpHNfJvQUHC5TNHcNt54xmfGce5kzJZcO1Mxgzry+atjk3vuJatewXKC2Hjm7D2FdutppQaEL0Zo2kbYDgTeNLZxdILVQpsCVFh3HnhJG5cuJzNxTXERYRwxwUTyE/vxwenH2puaaWmsZnY8FCC+rNJGZAeF8FVR+Vy4dRswoKDCAnu506Yece57rA54zuw4CLYt8a+jkyEK5+HzMn9ey+l1CH1JtC8JCLrgDrg+84CznrvVCuwTciOZ9F3jmJPRT1xkaEMH2StmfV7qnjo460s2VrKKYelcen0HEamxvT7ulFhvfnveBD5p8Gq52Dnp/b18JlQW9IeZMCuhfnsbjjvbgj20PsqpdzqzWSAm0TkT0ClMaZFRGqB89p+LiKnGmPe8kYlA5FHd4z0I3sq6vj2I18cyFTw3w+3snxHOQ9cPY34yDAf186RlAdzHrdbORsDqWPhw7tcjytaAU21EKxbMCjlTb3qozDGlBljWpznNcaYPR1+/CeP1kz5pS3FNZ3S4QAs3V7G9pJaH9WoG9EpMOJoyD3GPs89zvWYCZfoPj9KDYB+doZ3ogM2Q0BoiOt/GREI7e+4irflzLCbyIVGggTB5Mth4kW+rpVSQ4InO6e73bNGDR6jU2M4cWwq764vPlB2xYwc8lL8fBwqKgmOuRHGnw+tLXYtTMjg69pUyh/pKKjqlcToMH53/gS+2FbC6t2VHJGTyJG5iUSEBsB/paAgSBp56OOUUh7lyU+HbR681tBSW2a/XYf5eavAkZUYSVZiNucf4euaKKUCQU+2cv7GwX5ujHnW+fOgxyk3Kotg1f9g6QMQmwUn3gQ5R9tv3kopNUj0pEVzzkF+ZoBnPVSXocUYWP643ecEoHQLPPopzHsbsrSpoJQaPHqylfO3vPHGIvIj4NvYYLUS+BYQBSwEcrFdcRcbY8qc428G5gEtwPXGmDec8qnAw0Ak8CpwgzHGiEg48CgwFSgBLjHGbPPGvfRJ9V5YcnfnstZm2LtywALN7vI66ptayIiPJDJscGSDVkr5n16N0Tj70Iyn81bOt/X2TUUkC7geKDDG1InIImAOdpvod4wxfxSRm4CbgF+ISIHz8/FAJvC2iIxx1vTcA8wHPsMGmtOB17BBqcwYM1pE5mDX+VzS27p6TVAoRCS4bp4V6v1xmvqmZl5duYfbXl5DeW0TpxWkc/OZ48hL6f/qfqWU6qo32ZvvxX5Q/xC7ZuYiYEQ/3jsEiBSREGxLZjc208Ajzs8fAc53np8HPGWMaTDGbAU2AdNFJAOIM8Z8aowx2BZMx3ParvUMcLL4U3K26GQ45dbOZbGZA5J7a+WuSn686CvKa+1edm+u2ct/3t1MY7Nrun6llOqv3rRojjbGTBSRr40xt4rIX+nj+IwxZpeI3AnswOZOe9MY86aIpBtjipxjikQkzTklC9tiaVPolDU5z7uWt52z07lWs4hUAMlAp43eRWQ+tkVETo6bnRQ9pLKukWXby/lgQzG5KdEcl5/CyPxT4epXYPsnEJNuV7Inj/ZaHdpsKa52KXvpq938+NQxZCZEev39lVJDS28CTVvekVoRycSOe+T15U1FJBHb4sgDyoGnD7G3jbuWiDlI+cHO6VxgzH3AfQDTpk3z2qLT51fs5tcvrD7wOjsxkievncnw3GMh91hvva1biVGuOclGpsYQreM0Sikv6M082pdFJAH4C/AldrD+qT6+7ynAVmNMsTGmCdsyOhrY63SH4fy5zzm+EBje4fxsbFdbofO8a3mnc5zuuXigy4DIwNhdXsedb3TerriwrI41uyvdn9BUD/s3Quk2OzvNwyZmJ3Ds6Pb97sOCg/j12YcR7yYAKaVUf/Ume/PtztP/icjLQIQxpqKP77sDmCkiUdiW0snAUqAGmAv80fnzBef4F4EFInIXdjJAPvC5k0W6SkRmAkuAq4B/dThnLvApcCGw2BnHGXAtxlDf5Lo1caO77YrLtsHiO2DV0xASCSfeAlOuhAjPbeM8LD6Cv10ymbVFlVTVNzE6LYYxg2y/HKWU/+hxoBGRq9yUYYx5tLdvaoxZIiLPYFtGzcBybPdVDLBIROZhg9FFzvGrnZlpa5zjf9CWRRr4Hu3Tm19zHgAPAI+JyCZsS2ZOb+vpKZnxkVx9TC73fbDlQFl0WDDjuu4iaQx8+RisXGRfN9XCm7fYNPf5p3q0Tqmx4aTGpnr0mkop5Y709Eu+iPyrw8sIbCvkS2PMhd6omC9MmzbNLF261CvXLqqo47WVe1j4xU7y02O49riRTBqe0Pmg2lK4bxaUb+9cfvzP4aRbvFIvpZTqLxFZZoyZ1t3Pe9N19sMuF44HHutH3YaUjPhIrjk2j0uOzCYsONhtun3CoiH9cNdAo4kglVIBrD9JNWuxYyXKnbIdULkTIpPslOXgUACiw0O7PyckHI7/CWz7ABqqbFnGZBhxlPfrO9gUfQXbPgbTAiOOgcwj7MY5SqkB15sxmpdonx4chF3Fv8gblQp42z+FhZfbfeqDQ+G0O2DKVXbTrUPJmgrXvgv71tnj08dDXIb36zyY7PoSHj7LjnEBBIfZ9UrDp/u2XkoNUb1p0dzZ4XkzsN0YU9jdwUNW9T547rs2yAC0NMFrP4esaZA9tWfXSMm3D2+rr7QfxjHpg+vb/qpn24MMQEujzZCtgUYpn+jNGM373qzIoFFTDOXbXMsrC7H5PbtobbUJNkMjITLBy5Vre88W2PoBvP1bW68pV8PUqyFh+CFODBDVe1zLqvbav2vdgkGpAdebXGffEJGNIlIhIpXO+pVuVhwOYVEpEO/mAzsuy7WsfCcs/h3ccxQ8eDpsfMu2gLyt6Ct44ptQtAJq9sOHd8Ln99kP4sFgopvcqdOu0SCjlI/05jfvz8C5xph4Y0ycMSbWGBPnrYoFrNh0uODe9gWWEgSn/Q7SDut8XGsrLH0IPvor1JVB8VpYcLENAt62b41t1XS09EGo3AVFX8OWD6Bsu/tzA0HOUXDxo3YGX+pYuOA+yJvl61opNWT1ZoxmrzFmrddqMpjkHgvzP4CKHRCVDMn5ENIlvUv1Xlj2YOcy0wp7V0F2t9PRPSPMzXYAY2bDl4/Ch3+xC0cjE+GyhTB8hnfr4g3hMVBwHow8CTAQod+HlPKl3gSapSKyEHgeaGgrbNvKWXWRlGsf3QmJsIPwdWWdy8MH4EMxczKkjIH9G9rLCi6ARR3ymtaVwYs3wrdegagk79fJGyI0rY5S/qA3gSYOu3bmtA5lupWzG6U1DSzdVsbHm/czJi2WY/JTyE2O7nxQVCKcehs8eUl74syk0Xa9h7cl5sJli2DXMpuNIGOi+66y4jVQW9bzQFO9z06GiEqxXYhDVcUuqN1vv0jEDvN1bZTyud7MOvPKls6DTUur4bHPtvO3tzYeKCvIiOPBq6cxLL7LOpqRJ8I1b8KeVRAZD5lTIKlPOy/0XlJel/dyk4po2ES7QVtP7PjUTusu2wbx2XD+vZB3nCdqGjiMgc2L4bnv2IAblwnfuB9yj/F1zZTyqd4s2EwFrgVyO55njLnG89UKXDtLa7n73c2dytYUVbJuT5VroAkJs2s7/GF9R/rhtoX1zq12okDsMDjn7z2bcl2+E566vH3tUEWhXbA6//2BC5z+oGQzLLyifQ1P5W5YdKX9exgsU8eV6oPedJ29AHwIvA3onr/daG5pdpv+v7mx3ge16YXwWJjxfZsluq4CEkfYb+Q9UbGzPci0qa+A8h1DK9BU7Oy8UBTs30tFoQYaNaT1JtBEGWN+4bWaDBLDgys4/7A4nlvTvsQoMSqU/Ki6g5zlJ0JCIa2g9+dFJtpUOx3XAEmQnXE3lESn2vs2Hb5ohERAdEr35yg1BPR2h80zvVaTQSKcRn6c+gU/PSqW/LQYLhwfy+Mn1TOidRBn60nOh9l/6Fx26q22fChJzrd57dqIwFl3QdIo39VJKT/Qm/1oqoBo7NTmJkAAM5gWbXpkP5rmJnjnt7Dqf1RnHUtE+UZC6stsUsfB3H3SVAf71trxmvgsu0A1LPrQ5w02TXVQvA4qi+ykiNRxrmuoulFe22hnK27az8i0aI4dlUJeqps1T0r5GU/uRxMrIknYrQEiPFG5QSkkFI7+IaQfTsyaF6HgHDjsnMEdZMDmasuaYh+H0lhjc4+Fx0LMINvlMzTSTlHv5TR1YwyLlhby+1fb10SPSo3m0XnTyUqI8nQtlRpQvZl19m3gBiAbWAHMBD7B7rSpOoodBpMvtY8eqKhtJDwkmIiwYC9XzDPqm5rZuLeavZX1ZCZEMjotljB3G7m5s28dvPVr2PgGJOTC2X+107yDAuPevaWwrI6/v72hU9nm4hrWFVVpoFEBrzeTAW4AjgQ+M8acKCLjgFu9U62hYW9lPS99tZvHP9tOZkIk15+cz/TcJIKC+piyv7UVdn0By5+wWwBMudLm/Qrz3AdVQ1MLj326gzucb95BAn+9aBLnH5GFHGqrgYYqePWnsO1D+7p8Gyy4BOa/B8MmeKyOgail1dDQ7DpbscnNDEalAk1vJgPUG2PqAUQk3BizDhjrnWoNDU8v3cnvXlnLtpJaPtlcwpUPLGF1UUWfr2d2LYOHzoQvH4E1z8Hj36B16wcerDFsKq7m96+1d++0Gvjlc6vYtr/m0CdX7m4PMgcu0GzXnwxxWQmRXHpk5+7VuIgQxqRrGh0V+HrToikUkQRsrrO3RKQM2O2NSg0FeyvreeCjrZ3KmloMa3ZVMiEroU/XbF73GqGtzZ0LP/knjDrRbhPtAfurG+g6f6SuqYWy2iYOuWImLNqms6kt7VwemeiRugWy0JAgvn/iaIYnRfG/LwspyIhj3rF5jNTJAGoQ6M1kgAucp78VkXeBeOB1r9RqCAgNDiIuMpSy2s77z0T2Y5ymqaWF0C5lprWVxuZmwjwUaLIToggPCerUzZMaE86w+B7MD4nPhjPuhGfnted3KzjfbletyEyI5DsnjOKKmTmEhwQTEqz756jBoU//k40x7xtjXjTGNHq6QkNFUnQYN50xrlPZsPhwJmQn9Pmae9JnQVDn7w6Fh11LbWvX8NN3eSnR3HPFFJKi7ZTdjPgI/nP5EWQmRB7iTMe4s+Hbi+0eMVc8C2feqQsau4gOD9UgowaVHq+jGQo8so6mF+qbmlm+s5zPNpeSFhvOzJHJjErre1fJq1/vJKNqNaMLnyO0qYqtIy7iq6DDmFMQCYVf2PQo6YdDxqR+z/LaXV5HSU0D6bERpMXpbHelhjKPraNRnhcRGsJRI1M4aqRnvtFPzU3l3veH89meKwgNFoa1RvDLk+Pg4dl2EzawLZ4rnoWRJ/TrvTITInveilFKDWkaaAaR9LgIfnnGYWzdX0OzaSU3OZqo1U+1Bxmws7ze/T1kTbU7UQ5mTfWAsYsolVI+47OOYBFJEJFnRGSdiKwVkaNEJElE3hKRjc6fiR2Ov1lENonIehGZ3aF8qoisdH72T3EWc4hIuIgsdMqXiEiuD25zwIWGBDFmWCwFGfFEhYVATYnrQZW7oNnPs0n3R1MdbHgDHjsfHjkX1r4EDdUef5v9VQ28t34fi5bu5PMtJdQ0NB/6JKWGIF+2aP4BvG6MuVBEwoAo4JfAO8aYP4rITcBNwC9EpACYA4wHMoG3RWSMMaYFuAeYD3wGvAqcDrwGzAPKjDGjRWQO8CfgkoG9xXZNLa2sK6pkc3E1CVFhFGTEDczYxvAjXcumzRvcA/A7lsCCi9tfL7wCLn0Kxp7hsbcor23ktpfX8OJX7TP8f3tuAVfNzO3Rgtvy2kaW7yxnVWEFeSnRTBmRqF2RatDySaARkTjgeOBqAGf2WqOInAfMcg57BHgP+AVwHvCUMaYB2Coim4DpIrINiDPGfOpc91HgfGygOQ/4rXOtZ4B/i4gYH81+eH99MfMfW0qr8+7H56dw50WTvB9sMqfAJY/DW7+BulKY8V2Y5LN4OzC+Xuha9sX9kD8bgjzTiN+wt6pTkAH442vrmDUmjdyUgycTbW5p5dFPt3FXh11YjxmVzD8vPYLkGM9MQ1c+UlkEjdV2L6ehmFS2G75q0YwEioGHRGQSsAyb4ibdGFMEYIwpEpE05/gsbIulTaFT1uQ871reds5O51rNIlIBJAP7O1ZEROZjW0Tk5OR46v46Ka6q5/+9sOpAkAH4YON+Vu+u9H6gCY2wST1HHAPNDTYP26FSxQS6CDcJxcMTPBZkAKrqXbvJ6ptaqW08dPfZ9tJa/r24czaEjzeXsGFvFUdpoAlMzY2w4XV49SdQvQ9Gnwqz74BUTZ4CvhujCQGmAPcYY44AarDdZN1x98loDlJ+sHM6FxhznzFmmjFmWmqqdzIJ1za2sKeynuzECOZPjeW0/FiCBCrqBnAZUlQSxGUM/iADMOEiCO6Qmj8oGKbP8+hb5KVEE9Vlce2k4fFkJR66+6uxudXtLqz1TZrXLGDtXQVPX2WDDMCmt+Ct30JjAGx4OAB81aIpBAqNMUuc189gA81eEclwWjMZwL4Ox3dMBJWNTX9T6DzvWt7xnEIRCcFmMuiS+2RgpMeF899z05hU/DKpm56mIT6PDedcT3B6LM0treytaiAiJMhr3SaNzS1sK6mlobmFnKQo4iN7tj9KwMqaCte8DhvfgpZmGHOaLfOgkakxPHrNdG57eTVrdldx8rg0fjp7bI/+bocnRnJ8fgofbGxvXCdEhTIyVbtaAlbJJlxyM214Far3DK3tzLvhk0BjjNkjIjtFZKwxZj12q4E1zmMu8EfnzxecU14EFojIXdjJAPnA58aYFhGpEpGZwBLgKuBfHc6ZC3wKXAgs9tX4TIQYTtz/JMFfPgBAeOUuJhQtpXbum/zu5RCe/GIHqbHh/PrsAmaNTSUsxHMp88tqGrn/oy3c+/4WWloN00Yk8qdvTuzXwlC/J2IDi4eDS1fTcpN4bN5MquqbSIkJJyK0Z/9uMRGh3HreeB77dDuvrtzDpOx4rjs5nxHJGmgCVmSSa1n8cB2ncfgsM4CITAbuB8KALcC3sF15i4AcYAdwkTGm1Dn+FuAaoBm40RjzmlM+DXgYiMROAvihMcaISATwGHAEtiUzxxiz5WB18lpmgLJt8K+pdg1LB8Wz7+bIFxIOvBaBZ757NFNHeC7J5OJ1e7nm4c73dOXMHH5zzvjASHPS0mLHVgZhl19Lq6G8tpGY8BDCexiklJ+q2gsv3wjrX7Wvg4JhzlO2NT0E+G1mAGPMCsBdxdxupGaMuQO4w035UuBwN+X1wEX9q6WHBIfZAeouWYsbu6TANAbW76lkaipQtgVCIux+86HuJwyU1TSyZX81xsDI1GiSol273tbsrnQpe3PNXm44ZQwp/jzwXFsKm9+BpQ/Z3UmPvBayu/1/HJCCg0RnmQ0Wselwzj9h+nw7uzM5X5PFdqCZAQZCaDTM/AEsvr29LHk0YXHpQOeFhLOSy2HBPJubTASmfxeO+zHEpHU6bntJDT99+iu+2FYGwJScBO66eLLL1Fp3aean5iQSG+Hn//Srn4NXfmyfbwdWPw/z3oKMib6slVLdi0mFmBN9XQu/5OefNoNEbQlsfBNO/g1U7YaIeGhuILpuFxdNHU9Ggm2xVNc1kL7+MRtkwDZxltwDecfBuLM6XfLttfsOBBmAL3eU8/rqPXz3hFGdjpuSk8hJ49JYvM7Oq0iJCeMHJ44m3IPjQB5XXQwf/rVzWXM97P5SA41SAUgDzUCISgbTCu/cChEJ0FQDLU3snLuSrz7cytPLqgkOEr5/Qh7V1c3Edz1/93KXQPPhhmKXt3l//T6XQDMsPoK/XjSJjfuqqW9qYWRqNNmJfr4HvQRBsJutDYL0v6tSgSgARoMHgch4OPMvEJcF9eVgWmk477/84+NiNuy1XWctrYZ/vbuFr7Mvcz0/dZxL0Ynj0lzKTj4s3e3bJ0aHMT0viePHpPp/kAGIToZZN3cui4i3WQ6UUgFHvyIOlMwj4NrFdgZaRByloSN4/7kPXQ7b1hjLcfE57RmXR86C2EyX404al8Z764t5d73tEjsuP4XTCoZ58QYG2Liz4LJFsPpZO030sHMhvaB316jcZbsf47L8b9ZaSxMUr7f/H2LSIO0wCI/1da2U8goNNN7S3AS1xfbDo+0DJHaYfQBxjc1MyUnkw02dMuKQLcUw7gz7DV6C7Yrjre9D7tGdjhueFMU/5kxma0kNGMhNiSY+0nM7afZafSXsWQnlO+yWzRkT7T30VXgsjJltH71VWwpfPwXv/RFaW+xkiiOudJlQ4VPrXoH/XWPrB3Dsj209NdioQUgDjTfs3wQf/w3WvACph8Fpt0POzE6HRIeF8PPTx7LmoUpKamwqmgunZjGh8WNY8n+drzf6NDuhoGSzHadIHg0RccRFhjKpH1s/e0xzI3z+f7D4d+1lx/0Ejv+Zb/aC2fYRvN6h6+2d2yA2Aya76Zb0hbLt8NIN7UEG4KO7YNyZkO0m27ZSAU4Djac1VMNrv4DNb9vXhZ/DYxfA/PdcEuxNyE7gheuOYdv+GqLDQxidFkPs3jr4JNwmwARIzIOMCfDEJbDLmY027mw440+25eAPSjbCe3/oXPbRXVBwnt02eqCt+p9r2fLHYeKcHifW3F1eR11TCxnxEXZfH0+qL7ePrqpdJ3gMlOaWVnaU1tLUYhieFOn5e1ZDmv5v8rSKwvYg06apFvZvdJvJNTsxqvMAfc5M+PY7sHeNbQ1kHQHLHmsPMgDrXoaxZ8IRl3vpJnqpvrLzt3OwYyP1Fb6pj7uMuWkFPQoy9U0tvLaqiFtfWkN5bRMnjUvllrMKGOVmPVKfxWZC0kgo7ZCoIigEEnM99x69UFbTyCOfbuPudzfT2NLK6eOHcfOZ4zQljvIYnXXmaaEREObmQ6mn2yaLwLAJds+YgnMhKgU2vu563LaP+ldPT0ocYfff6Cg6xbbGfKHgPDulvE1EvB2j6YHVuyv40cKvKK9tAmDxumL+/vYGGppboKEKqvbYRJ39EZMK33zArh4HW9eLH4WUMf27bh8t217G39/eeCCj9Our9/D00kJ8lZ5KDT7aovG0hBFwyq12X4o2I0+y36j7IjTS7m1R9FXn8hHHuD++tsTOZmppsh9ccRl9e9/eiMuESxbAaz+zi00zjoCz7rSpY3whfTxc8ybsXQmtrTDs8B7vC7KluMalbMnWUmTHZ/DurbB/Axx+Ecz8HiSP7Hsds6bYDNNVRXZtla/+roCl212Tmr/09W6+fVweCVGDPNO3GhAaaDxNxLZGUsdC8Tr7QZ85xc4m2v6JnZkVnQbZUyGhhxutTZpj837tXm5f58+GkcdDU73NiWawqchriuH562Db+/a4pJEwZ4GdOuttWUfA5f+zeZ4iEyEywfvveTApo+2jl5KiXT9Y7zwhjLAF37DZCQC+uA9q98P593Sbh65HolP8Ykvt0W4yeU8enkC0jtMoD9H/Sd4QHmvTxuQd11729SJ49tr21xmTbRCIz3I53UVKPlz+tJ3NFhxiu1yaauHVn8GKx+wxE+dAwfntQQbsGMCyR2D27z26u2S3IuPtY6BV77NjYBJkW3HRyYc+pxuHZ8Uza0wq7zmZF8KCg5gata89yLRZ8xyceEufgpm/mZGXzJScBL7cUQ5AYlQo1x43ktAQ7VlXnqGBxkMaW1oIC+4mf1hlEbzRZaV70QrbHdaTQAMQnWofbda8AMsfbX/91ZO2BRMR33kQfuv7Nij1dIwo0OzfCIvmwr7V9vWIY+C8uyEpt0+XS4+L4M6LJrJuTxUVdU2MSo0hqnqJ64Hh8RAyODIvD0+K4v+unMr6PdU0NLeQnxZDjk4EUB6kgaafdpbW8urKIl5btYejRiXzzSOyGJ3eZdFdcwPUlbme3Og6HtDpHAm2LRh31r7sWrb1A9tNtuOz9rJxZw/eIAPw9cL2IAOw/WPY9DZM/3aPL1Fa3UBDcyvpcREEBQkpsREcG9uhS6xqAgyfCTs7/L2eertPx1U8LTU2gtTYfnQDKnUQGmj6oaa+mdtfWsOba/cCsGJnOW+s2sOT184kPb7DL21cJky6vHMLJDgM0lxzmFFXDpvesQsgo1PhqO/bD7mgLq2lEUfBpje7lB1tx212LrHTi3OPh0mX9vyGGmtsC6Gu3M4k8/ctaFuaYPNi1/Ltn/Yo0DQ0tbB4/T7ueGUtZTWNXHnUCK46KpfMhC6LTGOH2Vliu5baWWfp4zXvmlK9oIGmH7aV1hwIMm227K9hU3F150ATEgbH/8RufvbVAkgcCafeCuku+7XZHfqe/1776w2vwzVvuG76ddi58NVTsH+9fZ00Cg6/0C7inHgxtDbZ9+npmEldJXzy9/b0/BHxcNlCyDmqZ+f7QnAojDsHdi3rXD5qVo9O/6qwgu89/uWB1/e+v4WI0GBuODkf6ZobLSHbPpRSvaaBph+CRRCxjYeOgtwlcEzMtd0tR19nN0KLiHM9pr4SPv5757LWZrtmpmugSRkNV71gZ7YZY1tHbWtZhrkJYIeyd2XnPWDqK+DFG+Ga1yDKzX7oA2BnaS1b99cQFR7M6NQY91Ntx59vx6G2vOu8/gaM6tnmU18XlruUPbFkB5fPGEFq7OAYf1HKH2ig6YcRyVFcPDWbhUsLD5RNHp5Afno3YyJBQTbnVnckCELc5AZzM+jc2NzKsuIwnvw8nlYDl80IY2pki9u952sbmllTVEl9cwujU2MYFu/mPSp3u5btXwe1ZT4JNCsLy5n70BeUOnngzp6Ywa/PLiAtrss4QlIeXPSI3VBOgmym5h6OSSW7mcqclRBBZKjOtlLKkzTQ9ENkWAg/OnUs0/OSeH9DMVNGJHHi2FRS+roPfHgMnPBzeKpD8sewGMh1XZz55fYyLrv/swOtqZe/LmLBtTM4elTndRmFpbXc/f5mnvx8B8ZAQUYsf/jGRCYNT+h8QXcD2xlH9GuqcF/VNbZw55sbDgQZsPd33uQsTi3oEmgaa+yY1JJ77LjXzO/b7r6QQy80nJKTSF5KNFv320kZIUHCz2aPIybCh1mwlRqENND007D4CL45dTjfnOqhGUijTrJdYmtftulnxp4Ow1y3L3562U6XLrsnPtvuEmiW7ShjwZIdB16vKarioY+38qdvTuzc+kmfYNfbvP0bO8gelwXn/M39wsuSTbZ+W96zOdfGnt7zxac9UFnfxIqd5S7lO0trXQ/e9jEsuKj99cY3YO7LkHvsId9nREo0j3xrOqt2VVDT2MzYYbGMz/TBOiClBjkNNP4mNNJudjZy1kEPc7eNl8sANrCuqNKl7NMtJeytrO+8ViI8GqZ/B0afDHUVkJjjvpuveh88M8+uAwI7NrL1A7jgHo/tpZIYFcpJ49J4bvmuTuUuK9hbmm1LpiNjbPbmHgQagJzkKHKSA2DXUaUCmHZGB6iLpg3vtGmkCFw+w7VVMTLFdbxiUnaC2/EJgkPsttE5M7ofS9q/oT3ItFn3EpRscXt4X4SFBPP9WaOYmG0nTIQECTeenM/E7C6tDZFuxrS8sAdO6TZYsQDe/b2dfl7nGsCVUu5piyZATRmRyFPXzmThUtuFdvG04UwdkeBy3OScBE4al8bidXbL54z4CL59XB7RfR6H6GZLZA/vlJyfHsuj18xgR2ktkaHB5KZEExrc5XtRUDDM/C5seLV96l9wKIy/wLOVqdgFi66weeranPFnmPEdz76PUoOUaCrwdtOmTTNLly71dTU8bldZLev3VFHX1MKo1BjGZbiZWt1T1fvgiYuhaHl72bhzPNp11ivNjTZj9KpnITQcCi6ArKmeze224Q1YcHHnsvBY+N4nHh2bUipQicgyY8y07n6uLZohICsxiqxED41DxKTBhffbyQBb34cxZ8KY2b7b6z4kzM7KczMzz2Oa6u2kh8zJdqJEcBgsfwKam7z3nn1QUdfE9pIaQoODyEuJIiJUf72Vf/Dp/0QRCQaWAruMMWeLSBKwEMgFtgEXG2PKnGNvBuYBLcD1xpg3nPKpwMNAJPAqcIMxxohIOPAoMBUoAS4xxmzzxn1U1DVRUt1AfGQoyX2d2txXbS1Sd4tEvSV5NBx7o30MBSn5gNjxGbBddmf9za8yBWzdX8PNz37NZ1tKEYHLpudwwyn5pGn+MuUHfD0Z4AZgbYfXNwHvGGPygXec14hIATAHGA+cDtztBCmAe4D5QL7zON0pnweUGWNGA38D/uSNG1i1q4KrHlzCSX99n2/c8wmfbS5xf2BtGax7BZ7/AXz0d7uivz+aG2HL+7DoKlh4BWx+137z7oXG5hYq63v/rXx/VQPvrd/Hk0t28PGmYirqGg99UiCr2Q/rX2l/3doC7/2BlppSNuyt4rMt+9lecpAEqV5mjOHppTv5bEup89pmOPhiq+uGZkr5gs9aNCKSDZwF3AH82Ck+D5jlPH8EeA/4hVP+lDGmAdgqIpuA6SKyDYgzxnzqXPNR4HzgNeec3zrXegb4t4iI8eCg1P6qBn6w4Eu2l9j1HdtLarnmkS94+YfHMrLrHvNfPdl5q4Al/wfferXviSt3LoHHzmtv0ax72a6/OcS0aLAfTEu3l/GfxZvYWVbL5TNGcNbEYaTHHXq2VlV9E395Yz0Ll+48UHbjKfn8YNbowbt/SW2xa1lVETt27+Gsx4toajHERYRw9xVTOXb0wG9kVt3QzBur97qUL91WxlkTM92codTA8uUnw9+BnwOtHcrSjTFFAM6faU55FrCzw3GFTlmW87xreadzjDHNQAXgssxdROaLyFIRWVpc7OYD5SB2ldcdCDJtahtb2NF1YWHFTnjv953LqnZ3nsXUW8sfd02ytvShHp26encll/93Ce9tKGZzcQ23vbyGp77Y2aM94jftq+4UZAD+tXgTW334jd7rkka5dE225hzNP7+opqnF/p1V1jfzo4Ur2FNRN+DViwoL4ehRrmmCDs/qx6QPpTzIJ4FGRM4G9hljlh3yYOcUN2XmIOUHO6dzgTH3GWOmGWOmpaamujmle7ERIYS7+RafENVl6nCrsYPIXbX2YzC567YBAEE9a6Cu31NFY0trp7IHPtzKnspDd71VNzS7lLW0GmrdlA8aaQU2n1qU8z0l+0hKT/g9L6ztvJamuKqB/dUD340YHCRcMXMEOUntLdLj8lOYOXLg0wepALRvLax8BlY/ByWbvfIWvuo6OwY4V0TOBCKAOBF5HNgrIhnGmCIRyQD2OccXAh1zvGQDu53ybDflHc8pFJEQIB7waKd1bnI0f/zmBHaU1tHU0kpYcBARoUGMTusyAys+C466Dj68s70sIt79NgE9dcTltjvOOAFDBKZ9q0enuguO0eEhhPZgSnBeSjRJ0WGd8pCNTosmJ8m/VtdvKa5mze5KWo2hIDPO9d+kN0LCoOA8O226oRriMtlTYjB0TkSaFhtOSsyhc6x5w9hhcSz6zlFsLq4hLDiI0ekxJLrLdq1UR7u+hEfOgcZq+zom3XbBpx3m0bfxSaAxxtwM3AwgIrOAnxpjrhCRvwBzgT86f77gnPIisEBE7gIysYP+nxtjWkSkSkRmAkuAq4B/dThnLvApcCGw2JPjMwBBQUJwkPDPdzbS0mov/fPTx+LyOR4UDEdea/eKWf4opB0OR86D1LF9f/Ps6XD1K3ZPGtMKk+bYsh4YnxXHsPhw9lQ0HCi76YxxpPQgNX52YhQPf+tI/vDaWpbvKOf40an8ZPYYkgZ6tt1BrCuq5LL7lxwIhnERISy4dgaHZyX078Lx7d9p8tNa+N35h/PbF1cfGKO565LJ7jNjD5Bh8ZE+fX8VYFpb4PP72oMMQPVeu25sMASag/gjsEhE5gE7gIsAjDGrRWQRsAZoBn5gjGlxzvke7dObX3MeAA8AjzkTB0qxs9Y8antJDb98dtWBIAPwlzfWc3x+KodndUmXEjfMtjgmXwZBof1fUBgcanfUHHF0r0/NS4nhiXkz+WTzfooq6jlmVAqTcxJ6fP7E7ATuv/JIKuqbSIwKIzLMTTeeD726qqhTi6uyvpmFX+zsf6DpIDw0mEumDWd6bhIlNQ1kxkd2zh2nlL9raYLi9a7lXug+83mgMca8h51dhjGmBDi5m+PuwM5Q61q+FHDpgzLG1OMEKm8pq21yGbMwxvbVd8vN3jK+MCothlFdk1T2QnRECNERPv/v49bmfa4TEzbsraa11RAU5Ln1RiHBQeSnx5KPjxarKtUfoREwZS7s/rJz+bizPP5W/vlJESCGxYeTHhfO3sr2wBIWHER2oh92X1Tvg5ZGiM30bHoWL6lpaGbp9lJeX7WXzPgITilI57Aeps45e1IGr6ws6lR28bThHg0yYCdBbNtfQ2ltAxnxkWR7KvuCUgNl7Bm2u+yTf9iMFyf9Pxjh+e3bNddZB33JdbZ0Wyk/fHI5RRX1JESF8pcLJ3HyuDSPf6j1WVMdrH8d3rwF6svhyPkw/dudxhv80QsrdnHDUysOvI6LDOGZ7x7NmPRDtx7Kaxt5fvku7np7A62tcN2Jo7lwWnbfN6Rzo7G5heeW7+bXL6yiobmVhKhQ7r18KjNH6UwvFWCMgcpdzg61fVt3dahcZxpoOuhrUs19lfXsrWogMSrU/77VbvsIHu7SFD7pV3D8z3xTnx4oq2nkvP987LIe6S8XTuSiaT3fYG5PRT0GQ4YXBshX76rg7H9/1GkpU3pcOC/+4FjS4zXtixpaDhVo/L8PJQCkxUUwISve/4IM2MzGXS172KZV8VOtxtDY3OpS3tTiWnYww+IjvBJkAHZX1Lmsl91b2UBx9UHG55QaojTQDHYxaa5l8SPsTp5+KjkmnO/NGtWpLDwkiInZCb6pkBvD4lxbLSkxYST7aB2NUv5MA81gN3wmJHbIpxYcCif9EsL8eyruOZMyuOviSRyRk8DZEzNYcO1Mxmf6T0qV/PRYbj23gGBnLC4qLJi7Lp7stRaUUoFMx2g6GKwbn1G2DXavsBMD0gsgfUJAzDwDO+geHBR04APdnzQ2t7Blfw0lVQ1kJkaRmxyFDOR2DUr5Cd34TEFirn0EoLAQ/1oM2lFYSDDjhsXBMF/XRCn/poFGKTXo1dQ3s720huCgIHJTogj34y8wg5EGGqXUoLa9pIbbXlrDO+v2ESRw1cxcvn/SKN19dAAFRke9UoGktRUaalz3C1I+8fzyXbyzziaCbzXw8Kfb+HyL7j46kDTQKOVJxevhzV/BAyfDO7fBfu/s76F6pqahmddW7XEp/6S7LdeVV2jXmVKeUrUXFl4J+52MuPvWwo5PYc6TEJXo27oNUZGhwUzPTWLdnqpO5ROy47s5Q3mDtmiU8pSSje1Bps2OT6F0i2/qowgKEi6fmUNWQvt4zNQRCRwzOsWHtRp6tEWjlKcEhbovD+6mXA2IscPieOa7R7OpuJqQoCDy02M8mmBVHZoGGqU8JXUM5J8GG99sL5t0GSSP6v4cNSAyEiLJSNCsDb6igUYpT4lMhLP+ajNm7/zC7n6ae4zfp/tRyts00CjlSQk5drvuyZf5uiZK+Q2dDKCUUsqrNNAopZTyKg00SimlvEoDjVJKKa/SQKOUUsqrNNAopZTyKt1hswMRKQa2D8BbpQD7B+B9vG2w3AfovfirwXIvg+U+wP29jDDGpHZ3ggYaHxCRpQfb9jRQDJb7AL0XfzVY7mWw3Af07V6060wppZRXaaBRSinlVRpofOM+X1fAQwbLfYDei78aLPcyWO4D+nAvOkajlFLKq7RFo5RSyqs00CillPIqDTReJCLDReRdEVkrIqtF5AanPElE3hKRjc6ffr+hvIhEiMjnIvKVcy+3OuUBdy8AIhIsIstF5GXndaDexzYRWSkiK0RkqVMWqPeSICLPiMg653fmqEC8FxEZ6/x7tD0qReTGAL2XHzm/76tE5Ennc6DX96GBxruagZ8YYw4DZgI/EJEC4CbgHWNMPvCO89rfNQAnGWMmAZOB00VkJoF5LwA3AGs7vA7U+wA40RgzucPahkC9l38ArxtjxgGTsP8+AXcvxpj1zr/HZGAqUAs8R4Ddi4hkAdcD04wxhwPBwBz6ch/GGH0M0AN4ATgVWA9kOGUZwHpf162X9xEFfAnMCMR7AbKdX5CTgJedsoC7D6eu24CULmUBdy9AHLAVZ4JSIN9Ll/qfBnwciPcCZAE7gSTsJpkvO/fT6/vQFs0AEZFc4AhgCZBujCkCcP5M82HVeszpbloB7APeMsYE6r38Hfg50NqhLBDvA8AAb4rIMhGZ75QF4r2MBIqBh5wuzftFJJrAvJeO5gBPOs8D6l6MMbuAO4EdQBFQYYx5kz7chwaaASAiMcD/gBuNMZW+rk9fGWNajO0OyAami8jhPq5Sr4nI2cA+Y8wyX9fFQ44xxkwBzsB2zR7v6wr1UQgwBbjHGHMEUIOfdy0dioiEAecCT/u6Ln3hjL2cB+QBmUC0iFzRl2tpoPEyEQnFBpknjDHPOsV7RSTD+XkGtoUQMIwx5cB7wOkE3r0cA5wrItuAp4CTRORxAu8+ADDG7Hb+3IcdB5hOYN5LIVDotJIBnsEGnkC8lzZnAF8aY/Y6rwPtXk4Bthpjio0xTcCzwNH04T400HiRiAjwALDWGHNXhx+9CMx1ns/Fjt34NRFJFZEE53kk9j/hOgLsXowxNxtjso0xudhujcXGmCsIsPsAEJFoEYlte47tP19FAN6LMWYPsFNExjpFJwNrCMB76eBS2rvNIPDuZQcwU0SinM+yk7ETNHp9H5oZwItE5FjgQ2Al7eMBv8SO0ywCcrD/mBcZY0p9UskeEpGJwCPYmSdBwCJjzG0ikkyA3UsbEZkF/NQYc3Yg3oeIjMS2YsB2PS0wxtwRiPcCICKTgfuBMGAL8C2c/2sE3r1EYQfSRxpjKpyygPt3cZYxXIKdQbsc+DYQQy/vQwONUkopr9KuM6WUUl6lgUYppZRXaaBRSinlVRpolFJKeZUGGqWUUl6lgUYppZRXaaBRSinlVRpolPITIvK8kxxzdVuCTBGZJyIbROQ9EfmviPzbKU8Vkf+JyBfO4xjf1l6p7umCTaX8hIgkGWNKnRQ/XwCzgY+xOb+qgMXAV8aY60RkAXC3MeYjEckB3jB23yOl/E6IryuglDrgehG5wHk+HLgSeL8tvYeIPA2McX5+ClBgU1ABECciscaYqoGssFI9oYFGKT/g5F07BTjKGFMrIu9hN5jqrpUS5BxbNyAVVKofdIxGKf8QD5Q5QWYcduvvKOAEEUkUkRDgmx2OfxO4ru2Fk5BSKb+kgUYp//A6ECIiXwO3A58Bu4DfY7N9v41Nm1/hHH89ME1EvhaRNcB3B77KSvWMTgZQyo+JSIwxptpp0TwHPGiMee5Q5ynlT7RFo5R/+62IrMBuaLYVeN6ntVGqD7RFo5RSyqu0RaOUUsqrNNAopZTyKg00SimlvEoDjVJKKa/SQKOUUsqr/j/Xkl38MsDHZwAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# explore correlation between age and annual salary\n",
    "sns.scatterplot(data=final_data, x=\"age\", y=\"annual_salary\", hue=\"gender\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "c7a42728",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# explore correlation between average purchase amount and annual salary\n",
    "sns.scatterplot(data=final_data, x=\"avg_purchase_amount\", y=\"annual_salary\", hue=\"gender\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "d3639b25",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# explore correlation between average balance and annual salary\n",
    "sns.scatterplot(data=final_data, x=\"avg_balance\", y=\"annual_salary\", hue=\"gender\")\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f7d75bc7",
   "metadata": {},
   "source": [
    "# Final Preprocessing"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "578732e4",
   "metadata": {},
   "outputs": [],
   "source": [
    "# remove customer_id\n",
    "final_data = final_data.drop(columns = ['customer_id'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "id": "cc9346f0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "array(['F', 'M'], dtype=object)"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "final_data['gender'].unique()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "d71a4fa1",
   "metadata": {},
   "source": [
    "'gender' is a categorical feature, which needs to be converted numeric before fitting the model. Techniques such as 'one-hot-encoding could be used in order to do so, but for the puprose of this lets encode 'F' as 0 and 'M' as 1."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "id": "544d4f57",
   "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>gender</th>\n",
       "      <th>age</th>\n",
       "      <th>annual_salary</th>\n",
       "      <th>avg_purchase_amount</th>\n",
       "      <th>min_purchase_amount</th>\n",
       "      <th>max_purchase_amount</th>\n",
       "      <th>avg_monthly_purchases</th>\n",
       "      <th>avg_balance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>53</td>\n",
       "      <td>50464.44</td>\n",
       "      <td>37.726250</td>\n",
       "      <td>6.48</td>\n",
       "      <td>226.82</td>\n",
       "      <td>16</td>\n",
       "      <td>2275.852055</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>21</td>\n",
       "      <td>100202.20</td>\n",
       "      <td>76.458077</td>\n",
       "      <td>5.64</td>\n",
       "      <td>2885.85</td>\n",
       "      <td>17</td>\n",
       "      <td>9829.929000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>1</td>\n",
       "      <td>28</td>\n",
       "      <td>45996.24</td>\n",
       "      <td>67.531385</td>\n",
       "      <td>4.09</td>\n",
       "      <td>1271.46</td>\n",
       "      <td>21</td>\n",
       "      <td>5699.212250</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>34</td>\n",
       "      <td>88992.28</td>\n",
       "      <td>51.128289</td>\n",
       "      <td>8.78</td>\n",
       "      <td>432.92</td>\n",
       "      <td>25</td>\n",
       "      <td>9032.841186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>34</td>\n",
       "      <td>109304.44</td>\n",
       "      <td>30.310491</td>\n",
       "      <td>3.66</td>\n",
       "      <td>390.67</td>\n",
       "      <td>54</td>\n",
       "      <td>22272.433755</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   gender  age  annual_salary  avg_purchase_amount  min_purchase_amount  \\\n",
       "0       0   53       50464.44            37.726250                 6.48   \n",
       "1       1   21      100202.20            76.458077                 5.64   \n",
       "2       1   28       45996.24            67.531385                 4.09   \n",
       "3       0   34       88992.28            51.128289                 8.78   \n",
       "4       0   34      109304.44            30.310491                 3.66   \n",
       "\n",
       "   max_purchase_amount  avg_monthly_purchases   avg_balance  \n",
       "0               226.82                     16   2275.852055  \n",
       "1              2885.85                     17   9829.929000  \n",
       "2              1271.46                     21   5699.212250  \n",
       "3               432.92                     25   9032.841186  \n",
       "4               390.67                     54  22272.433755  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# encode 'M' as 1 and 'F' as 0\n",
    "final_data['gender'] = final_data['gender'].replace({'F': 0, 'M': 1})\n",
    "final_data.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "842d76c3",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "gender                     int64\n",
       "age                        int64\n",
       "annual_salary            float64\n",
       "avg_purchase_amount      float64\n",
       "min_purchase_amount      float64\n",
       "max_purchase_amount      float64\n",
       "avg_monthly_purchases      int32\n",
       "avg_balance              float64\n",
       "dtype: object"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# verify data types of all variables\n",
    "final_data.dtypes"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "939dddbd",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "gender                   0\n",
       "age                      0\n",
       "annual_salary            0\n",
       "avg_purchase_amount      0\n",
       "min_purchase_amount      0\n",
       "max_purchase_amount      0\n",
       "avg_monthly_purchases    0\n",
       "avg_balance              0\n",
       "dtype: int64"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# make sure there are no missing values\n",
    "final_data.isnull().sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e28dea28",
   "metadata": {},
   "source": [
    "# Train-Test split"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "cd188479",
   "metadata": {},
   "outputs": [],
   "source": [
    "X = final_data.drop(columns = ['annual_salary'])\n",
    "y = final_data['annual_salary']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "id": "cf9fc45f",
   "metadata": {},
   "outputs": [],
   "source": [
    "X_train, X_test, y_train, y_test = train_test_split(X , y, test_size = 0.25, random_state = 3)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c09592de",
   "metadata": {},
   "source": [
    "# Fitting Linear Regression"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "id": "c5798e41",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "LinearRegression()"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model1 = LinearRegression()\n",
    "model1.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "id": "2c029e33",
   "metadata": {},
   "outputs": [],
   "source": [
    "model1_pred = model1.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "id": "2d6fcfa4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "20991.79353794082"
      ]
     },
     "execution_count": 27,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.mean_absolute_error(y_test, model1_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "id": "70fd1ae0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "25360.345616140752"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sqrt(metrics.mean_squared_error(y_test, model1_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1a33938b",
   "metadata": {},
   "source": [
    "# Fitting Decision Tree"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "id": "4b9c761e",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "DecisionTreeRegressor()"
      ]
     },
     "execution_count": 29,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "model2 = DecisionTreeRegressor()\n",
    "model2.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "id": "667bdf78",
   "metadata": {},
   "outputs": [],
   "source": [
    "model2_pred = model2.predict(X_test)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "id": "5f45fe91",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "18582.264000000003"
      ]
     },
     "execution_count": 31,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "metrics.mean_absolute_error(y_test, model2_pred)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "id": "52308f05",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "24173.840775448494"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "np.sqrt(metrics.mean_squared_error(y_test, model2_pred))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17100fa3",
   "metadata": {},
   "source": [
    "# Summary"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "11090e5a",
   "metadata": {},
   "source": [
    "| **Model** | **MAE** | **RMSE**\n",
    "|:----:|:----:|:----:|\n",
    "| **model1 (Linear Regression Model)** | 20991.79 | 25360.34\n",
    "| **model2 (Decision Tree Regressor)** | 18582.26 | 24173.84"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a260e1ec",
   "metadata": {},
   "source": [
    "Although Decision Tree performed better than Linear Regression, the MAE and RMSE for both the models are very high. It suggests that both the models are highly inaccurate and cannot be used by ANZ to segment customers into income brackets for reporting purpose. On hypertuning parameters the models could have performed slightly better but we cannot rely on it because the data we have is only for 100 customers. So, inorder to build more reliable models we need to come up with advance feature engineering and feature selection methods and more data to train and test the model.\n",
    "\n",
    "If the goal is to segment customers, why not we create discrete class labels for annual salary and perform classification as it is more likely to give better results. However, more research is to be done on this part."
   ]
  }
 ],
 "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": "3.8.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}