market_analysis.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "from scipy import stats\n",
    "import matplotlib.pyplot as plt\n",
    "from sqlalchemy import create_engine\n",
    "from scipy.stats import f_oneway"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "engine_txt = 'postgresql://postgres:password@localhost:5432/olist'\n",
    "engine = create_engine(engine_txt)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Total products bought by state"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "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>state</th>\n",
       "      <th>tot_order</th>\n",
       "      <th>cum_share</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>SP</td>\n",
       "      <td>114377</td>\n",
       "      <td>0.385857</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RJ</td>\n",
       "      <td>40110</td>\n",
       "      <td>0.521171</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MG</td>\n",
       "      <td>36091</td>\n",
       "      <td>0.642926</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>RS</td>\n",
       "      <td>16847</td>\n",
       "      <td>0.699760</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PR</td>\n",
       "      <td>15016</td>\n",
       "      <td>0.750417</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  state  tot_order  cum_share\n",
       "0    SP     114377   0.385857\n",
       "1    RJ      40110   0.521171\n",
       "2    MG      36091   0.642926\n",
       "3    RS      16847   0.699760\n",
       "4    PR      15016   0.750417"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = \"\"\"\n",
    "WITH orders_by_state AS(\n",
    "    SELECT\n",
    "        customers.customer_state as state,\n",
    "        SUM(payments.payment_installments) AS tot_order\n",
    "    FROM order_payments AS payments\n",
    "    JOIN orders ON orders.order_id = payments.order_id\n",
    "    JOIN customers ON customers.customer_id = orders.customer_id\n",
    "    GROUP BY customers.customer_state\n",
    "    ORDER BY SUM(payments.payment_installments) DESC\n",
    ")\n",
    "\n",
    "SELECT\n",
    "    *,\n",
    "    SUM(tot_order) OVER (ORDER BY tot_order DESC) / (SELECT SUM(tot_order)\n",
    "                                                    FROM orders_by_state) AS cum_share\n",
    "FROM orders_by_state\n",
    "\"\"\"\n",
    "\n",
    "orders_by_state = pd.read_sql(query, engine)\n",
    "orders_by_state.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "data = {\n",
    "    'states': [\n",
    "        'São Paulo', 'Rio de Janeiro', 'Minas Gerais', 'Paraná', 'Rio Grande do Sul',\n",
    "        'Santa Catarina', 'Bahia', 'Distrito Federal', 'Goiás', 'Pará', 'Pernambuco',\n",
    "        'Mato Grosso', 'Ceará', 'Espírito Santo', 'Mato Grosso do Sul', 'Amazonas',\n",
    "        'Maranhão', 'Rio Grande do Norte', 'Paraíba', 'Alagoas', 'Piauí', 'Rondônia',\n",
    "        'Sergipe', 'Tocantins', 'Amapá', 'Acre', 'Roraima'\n",
    "    ],\n",
    "    'gdp': [\n",
    "        2377.64, 759.82, 682.79, 487.93, 470.94, 349.28, 305.32, 265.85, 224.13, 215.94,\n",
    "        193.31, 178.65, 166.92, 138.45, 122.63, 116.02, 106.92, 71.58, 70.29, 63.2, 56.39,\n",
    "        51.6, 45.41, 43.65, 18.47, 16.48, 16.02\n",
    "    ],\n",
    "    'iso_code': [\n",
    "        'SP', 'RJ', 'MG', 'PR', 'RS', 'SC', 'BA', 'DF', 'GO', 'PA', 'PE', 'MT', 'CE', 'ES',\n",
    "        'MS', 'AM', 'MA', 'RN', 'PB', 'AL', 'PI', 'RO', 'SE', 'TO', 'AP', 'AC', 'RR'\n",
    "    ]\n",
    "}\n",
    "\n",
    "info_states = pd.DataFrame(data)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "orders_by_state = pd.merge(orders_by_state, info_states, left_on='state', right_on='iso_code').drop(columns='iso_code')\n",
    "\n",
    "orders_by_state.to_csv(r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\tot_prod_bought_by_state.csv',\n",
    "                       index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>tot_order</th>\n",
       "      <th>gdp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>tot_order</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>0.992012</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>gdp</th>\n",
       "      <td>0.992012</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           tot_order       gdp\n",
       "tot_order   1.000000  0.992012\n",
       "gdp         0.992012  1.000000"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "orders_by_state[['tot_order', 'gdp']].corr()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>state</th>\n",
       "      <th>tot_order</th>\n",
       "      <th>cum_share</th>\n",
       "      <th>states</th>\n",
       "      <th>gdp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>SP</td>\n",
       "      <td>114377</td>\n",
       "      <td>0.385857</td>\n",
       "      <td>São Paulo</td>\n",
       "      <td>2377.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RJ</td>\n",
       "      <td>40110</td>\n",
       "      <td>0.521171</td>\n",
       "      <td>Rio de Janeiro</td>\n",
       "      <td>759.82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MG</td>\n",
       "      <td>36091</td>\n",
       "      <td>0.642926</td>\n",
       "      <td>Minas Gerais</td>\n",
       "      <td>682.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>RS</td>\n",
       "      <td>16847</td>\n",
       "      <td>0.699760</td>\n",
       "      <td>Rio Grande do Sul</td>\n",
       "      <td>470.94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PR</td>\n",
       "      <td>15016</td>\n",
       "      <td>0.750417</td>\n",
       "      <td>Paraná</td>\n",
       "      <td>487.93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>BA</td>\n",
       "      <td>11509</td>\n",
       "      <td>0.789244</td>\n",
       "      <td>Bahia</td>\n",
       "      <td>305.32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>SC</td>\n",
       "      <td>10760</td>\n",
       "      <td>0.825543</td>\n",
       "      <td>Santa Catarina</td>\n",
       "      <td>349.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ES</td>\n",
       "      <td>6301</td>\n",
       "      <td>0.846800</td>\n",
       "      <td>Espírito Santo</td>\n",
       "      <td>138.45</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  state  tot_order  cum_share             states      gdp\n",
       "0    SP     114377   0.385857          São Paulo  2377.64\n",
       "1    RJ      40110   0.521171     Rio de Janeiro   759.82\n",
       "2    MG      36091   0.642926       Minas Gerais   682.79\n",
       "3    RS      16847   0.699760  Rio Grande do Sul   470.94\n",
       "4    PR      15016   0.750417             Paraná   487.93\n",
       "5    BA      11509   0.789244              Bahia   305.32\n",
       "6    SC      10760   0.825543     Santa Catarina   349.28\n",
       "7    ES       6301   0.846800     Espírito Santo   138.45"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_8 = orders_by_state[orders_by_state['cum_share']<0.85]\n",
    "top_8"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## State with the highest order value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "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>order_value</th>\n",
       "      <th>state</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>20.24</td>\n",
       "      <td>SP</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>21.06</td>\n",
       "      <td>PR</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>9.66</td>\n",
       "      <td>SP</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>21.22</td>\n",
       "      <td>SP</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>9.55</td>\n",
       "      <td>PR</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   order_value state\n",
       "0        20.24    SP\n",
       "1        21.06    PR\n",
       "2         9.66    SP\n",
       "3        21.22    SP\n",
       "4         9.55    PR"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#ANOVA test\n",
    "\n",
    "query = \"\"\"\n",
    "SELECT\n",
    "    SUM(payment_value::NUMERIC) AS order_value,\n",
    "    customers.customer_state AS state\n",
    "FROM orders\n",
    "JOIN customers ON customers.customer_id = orders.customer_id\n",
    "JOIN order_payments AS payments ON payments.order_id = orders.order_id\n",
    "GROUP BY customers.customer_state, orders.order_id\n",
    "\"\"\"\n",
    "\n",
    "df = pd.read_sql(query, engine)\n",
    "df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "F_onewayResult(statistic=34.62863411435716, pvalue=3.149889058834275e-172)"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "states = []\n",
    "for state in df['state'].unique():\n",
    "    states.append(df.loc[df['state']==state, 'order_value'].values)\n",
    "\n",
    "f_oneway(*states)\n",
    "\n",
    "# The order value is influenced by the state"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "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>state</th>\n",
       "      <th>mean_order_value</th>\n",
       "      <th>std_order_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>PB</td>\n",
       "      <td>49.355168</td>\n",
       "      <td>73.505447</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>AC</td>\n",
       "      <td>45.410370</td>\n",
       "      <td>43.188524</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>RO</td>\n",
       "      <td>44.962806</td>\n",
       "      <td>57.541384</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>AP</td>\n",
       "      <td>44.697353</td>\n",
       "      <td>51.895554</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>AL</td>\n",
       "      <td>43.878232</td>\n",
       "      <td>51.838240</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  state  mean_order_value  std_order_value\n",
       "0    PB         49.355168        73.505447\n",
       "1    AC         45.410370        43.188524\n",
       "2    RO         44.962806        57.541384\n",
       "3    AP         44.697353        51.895554\n",
       "4    AL         43.878232        51.838240"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = \"\"\"\n",
    "WITH order_values AS(\n",
    "    SELECT\n",
    "        SUM(payment_value::NUMERIC) AS order_value,\n",
    "        customers.customer_state AS state\n",
    "    FROM orders\n",
    "    JOIN customers ON customers.customer_id = orders.customer_id\n",
    "    JOIN order_payments AS payments ON payments.order_id = orders.order_id\n",
    "    GROUP BY customers.customer_state, orders.order_id\n",
    ")\n",
    "\n",
    "SELECT\n",
    "    state,\n",
    "    AVG(order_value) as mean_order_value,\n",
    "    STDDEV(order_value) as std_order_value\n",
    "FROM order_values\n",
    "GROUP BY state\n",
    "ORDER BY AVG(order_value) DESC\n",
    "\"\"\"\n",
    "\n",
    "order_value_by_states = pd.read_sql(query, engine)\n",
    "order_value_by_states.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Is there a relation between avg_order_value and tot_orders by state?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>state</th>\n",
       "      <th>tot_order</th>\n",
       "      <th>cum_share</th>\n",
       "      <th>state_name</th>\n",
       "      <th>gdp</th>\n",
       "      <th>mean_order_value</th>\n",
       "      <th>std_order_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>SP</td>\n",
       "      <td>114377</td>\n",
       "      <td>0.385857</td>\n",
       "      <td>São Paulo</td>\n",
       "      <td>2377.64</td>\n",
       "      <td>26.852245</td>\n",
       "      <td>36.135222</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RJ</td>\n",
       "      <td>40110</td>\n",
       "      <td>0.521171</td>\n",
       "      <td>Rio de Janeiro</td>\n",
       "      <td>759.82</td>\n",
       "      <td>31.182029</td>\n",
       "      <td>45.748465</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MG</td>\n",
       "      <td>36091</td>\n",
       "      <td>0.642926</td>\n",
       "      <td>Minas Gerais</td>\n",
       "      <td>682.79</td>\n",
       "      <td>30.072543</td>\n",
       "      <td>39.704736</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>RS</td>\n",
       "      <td>16847</td>\n",
       "      <td>0.699760</td>\n",
       "      <td>Rio Grande do Sul</td>\n",
       "      <td>470.94</td>\n",
       "      <td>30.460115</td>\n",
       "      <td>37.665875</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PR</td>\n",
       "      <td>15016</td>\n",
       "      <td>0.750417</td>\n",
       "      <td>Paraná</td>\n",
       "      <td>487.93</td>\n",
       "      <td>30.047903</td>\n",
       "      <td>40.224941</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  state  tot_order  cum_share         state_name      gdp  mean_order_value  \\\n",
       "0    SP     114377   0.385857          São Paulo  2377.64         26.852245   \n",
       "1    RJ      40110   0.521171     Rio de Janeiro   759.82         31.182029   \n",
       "2    MG      36091   0.642926       Minas Gerais   682.79         30.072543   \n",
       "3    RS      16847   0.699760  Rio Grande do Sul   470.94         30.460115   \n",
       "4    PR      15016   0.750417             Paraná   487.93         30.047903   \n",
       "\n",
       "   std_order_value  \n",
       "0        36.135222  \n",
       "1        45.748465  \n",
       "2        39.704736  \n",
       "3        37.665875  \n",
       "4        40.224941  "
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_states = pd.merge(orders_by_state, order_value_by_states, on='state').rename(columns={'states':'state_name'})\n",
    "df_states.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_states.to_csv(r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\info_all_statea.csv',\n",
    "                 index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "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>gdp</th>\n",
       "      <th>mean_order_value</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>gdp</th>\n",
       "      <td>1.000000</td>\n",
       "      <td>-0.610035</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>mean_order_value</th>\n",
       "      <td>-0.610035</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                       gdp  mean_order_value\n",
       "gdp               1.000000         -0.610035\n",
       "mean_order_value -0.610035          1.000000"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_states[['gdp', 'mean_order_value']].corr()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Confidece intervals avg order value"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "metadata": {},
   "outputs": [],
   "source": [
    "def mean_conf_int(df, category, to_test, value, alpha=0.05, do_return=True):\n",
    "    '''Calculate the confidence interval for the difference between 2 means.\n",
    "    The Central Limit Theorem applies.'''\n",
    "\n",
    "    mean1 = df.loc[df[category]==to_test[0], value].mean()\n",
    "    mean2 = df.loc[df[category]==to_test[1], value].mean()\n",
    "    std1 = df.loc[df[category]==to_test[0], value].std()\n",
    "    std2 = df.loc[df[category]==to_test[1], value].std()\n",
    "    n1 = df[df[category]==to_test[0]].shape[0]\n",
    "    n2 = df[df[category]==to_test[1]].shape[0]\n",
    "\n",
    "    std_error = np.sqrt((std1**2 / n1) + (std2**2 / n2))\n",
    "    z_critical = stats.norm.ppf(1 - alpha / 2)\n",
    "    margin_of_error = z_critical * std_error\n",
    "\n",
    "    lower_bound = (mean1 - mean2) - margin_of_error\n",
    "    upper_bound = (mean1 - mean2) + margin_of_error\n",
    "\n",
    "    if do_return:\n",
    "        return lower_bound, upper_bound\n",
    "    else:\n",
    "        print(f'{to_test[0]} vs. {to_test[1]}')\n",
    "        print(f'Upper bound: {upper_bound} —— Lower bound: {lower_bound}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "conf_intervals = {'PB_vs':[], 'up_bound':[], 'lo_bound':[]}\n",
    "prev = 'PB'\n",
    "\n",
    "for state in order_value_by_states['state'].values:\n",
    "    if state != prev:\n",
    "        conf_intervals['PB_vs'].append(state)\n",
    "        curr = state\n",
    "        lo, hi = mean_conf_int(df, 'state', (prev, curr), 'order_value')\n",
    "        conf_intervals['lo_bound'].append(lo)\n",
    "        conf_intervals['up_bound'].append(hi)\n",
    "\n",
    "ci_order_value = pd.DataFrame(conf_intervals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 47,
   "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>PB_vs</th>\n",
       "      <th>up_bound</th>\n",
       "      <th>lo_bound</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>SP</td>\n",
       "      <td>28.735361</td>\n",
       "      <td>16.270485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MG</td>\n",
       "      <td>25.547098</td>\n",
       "      <td>13.018152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PR</td>\n",
       "      <td>25.628275</td>\n",
       "      <td>12.986255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ES</td>\n",
       "      <td>25.996509</td>\n",
       "      <td>12.784727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>RS</td>\n",
       "      <td>25.197448</td>\n",
       "      <td>12.592657</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  PB_vs   up_bound   lo_bound\n",
       "0    SP  28.735361  16.270485\n",
       "1    MG  25.547098  13.018152\n",
       "2    PR  25.628275  12.986255\n",
       "3    ES  25.996509  12.784727\n",
       "4    RS  25.197448  12.592657"
      ]
     },
     "execution_count": 47,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ci_order_value = ci_order_value.sort_values('lo_bound', ascending=False).reset_index(drop=True)\n",
    "ci_order_value.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "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>PB_vs</th>\n",
       "      <th>up_bound</th>\n",
       "      <th>lo_bound</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>SP</td>\n",
       "      <td>28.735361</td>\n",
       "      <td>16.270485</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MG</td>\n",
       "      <td>25.547098</td>\n",
       "      <td>13.018152</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>PR</td>\n",
       "      <td>25.628275</td>\n",
       "      <td>12.986255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ES</td>\n",
       "      <td>25.996509</td>\n",
       "      <td>12.784727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>RS</td>\n",
       "      <td>25.197448</td>\n",
       "      <td>12.592657</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>RJ</td>\n",
       "      <td>24.445993</td>\n",
       "      <td>11.900284</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>DF</td>\n",
       "      <td>24.848277</td>\n",
       "      <td>11.833432</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>SC</td>\n",
       "      <td>23.739358</td>\n",
       "      <td>10.936884</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>GO</td>\n",
       "      <td>23.484587</td>\n",
       "      <td>10.446046</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  PB_vs   up_bound   lo_bound\n",
       "0    SP  28.735361  16.270485\n",
       "1    MG  25.547098  13.018152\n",
       "2    PR  25.628275  12.986255\n",
       "3    ES  25.996509  12.784727\n",
       "4    RS  25.197448  12.592657\n",
       "5    RJ  24.445993  11.900284\n",
       "6    DF  24.848277  11.833432\n",
       "7    SC  23.739358  10.936884\n",
       "8    GO  23.484587  10.446046"
      ]
     },
     "execution_count": 51,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ci_order_value.head(9)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "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>state</th>\n",
       "      <th>tot_order</th>\n",
       "      <th>cum_share</th>\n",
       "      <th>states</th>\n",
       "      <th>gdp</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>SP</td>\n",
       "      <td>114377</td>\n",
       "      <td>0.385857</td>\n",
       "      <td>São Paulo</td>\n",
       "      <td>2377.64</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>RJ</td>\n",
       "      <td>40110</td>\n",
       "      <td>0.521171</td>\n",
       "      <td>Rio de Janeiro</td>\n",
       "      <td>759.82</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>MG</td>\n",
       "      <td>36091</td>\n",
       "      <td>0.642926</td>\n",
       "      <td>Minas Gerais</td>\n",
       "      <td>682.79</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>RS</td>\n",
       "      <td>16847</td>\n",
       "      <td>0.699760</td>\n",
       "      <td>Rio Grande do Sul</td>\n",
       "      <td>470.94</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>PR</td>\n",
       "      <td>15016</td>\n",
       "      <td>0.750417</td>\n",
       "      <td>Paraná</td>\n",
       "      <td>487.93</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>BA</td>\n",
       "      <td>11509</td>\n",
       "      <td>0.789244</td>\n",
       "      <td>Bahia</td>\n",
       "      <td>305.32</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>SC</td>\n",
       "      <td>10760</td>\n",
       "      <td>0.825543</td>\n",
       "      <td>Santa Catarina</td>\n",
       "      <td>349.28</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>ES</td>\n",
       "      <td>6301</td>\n",
       "      <td>0.846800</td>\n",
       "      <td>Espírito Santo</td>\n",
       "      <td>138.45</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  state  tot_order  cum_share             states      gdp\n",
       "0    SP     114377   0.385857          São Paulo  2377.64\n",
       "1    RJ      40110   0.521171     Rio de Janeiro   759.82\n",
       "2    MG      36091   0.642926       Minas Gerais   682.79\n",
       "3    RS      16847   0.699760  Rio Grande do Sul   470.94\n",
       "4    PR      15016   0.750417             Paraná   487.93\n",
       "5    BA      11509   0.789244              Bahia   305.32\n",
       "6    SC      10760   0.825543     Santa Catarina   349.28\n",
       "7    ES       6301   0.846800     Espírito Santo   138.45"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "top_8"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.barh(ci_order_value['PB_vs'], ci_order_value['up_bound']-ci_order_value['lo_bound'], left=ci_order_value['lo_bound'])\n",
    "plt.title('Conf. Int. Avg Order Value by State')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "metadata": {},
   "outputs": [],
   "source": [
    "ci_order_value.to_csv(r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\order_value_vs_state.csv',\n",
    "                 index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## State with the most recurring customers"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "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>state</th>\n",
       "      <th>is_recurring</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>MA</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>CE</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>ES</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>MG</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>MG</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  state  is_recurring\n",
       "0    MA             0\n",
       "1    CE             0\n",
       "2    ES             0\n",
       "3    MG             0\n",
       "4    MG             0"
      ]
     },
     "execution_count": 61,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# ANOVA test\n",
    "\n",
    "query = '''\n",
    "SELECT\n",
    "\tcustomers.customer_state AS state,\n",
    "\tCASE\n",
    "\t\tWHEN COUNT(*)>1 THEN 1\n",
    "\t\tELSE 0\n",
    "\tEND AS is_recurring\n",
    "FROM orders\n",
    "JOIN customers ON customers.customer_id = orders.customer_id\n",
    "GROUP BY customers.customer_state, customers.customer_unique_id\n",
    "'''\n",
    "\n",
    "anova_t_recur = pd.read_sql(query, engine)\n",
    "anova_t_recur.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "F_onewayResult(statistic=1.0598090762110413, pvalue=0.380720876490711)"
      ]
     },
     "execution_count": 62,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "states = []\n",
    "for state in anova_t_recur['state'].unique():\n",
    "    states.append(anova_t_recur.loc[df['state']==state, 'is_recurring'].values)\n",
    "\n",
    "f_oneway(*states)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Confidence intervals recurring customers proportion"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "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>state</th>\n",
       "      <th>size</th>\n",
       "      <th>recurring_size</th>\n",
       "      <th>pi</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>RJ</td>\n",
       "      <td>12384</td>\n",
       "      <td>421</td>\n",
       "      <td>0.0340</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>MT</td>\n",
       "      <td>876</td>\n",
       "      <td>29</td>\n",
       "      <td>0.0331</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>GO</td>\n",
       "      <td>1952</td>\n",
       "      <td>64</td>\n",
       "      <td>0.0328</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>SP</td>\n",
       "      <td>40302</td>\n",
       "      <td>1296</td>\n",
       "      <td>0.0322</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>RS</td>\n",
       "      <td>5277</td>\n",
       "      <td>167</td>\n",
       "      <td>0.0316</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  state   size  recurring_size      pi\n",
       "0    RJ  12384             421  0.0340\n",
       "1    MT    876              29  0.0331\n",
       "2    GO   1952              64  0.0328\n",
       "3    SP  40302            1296  0.0322\n",
       "4    RS   5277             167  0.0316"
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = '''\n",
    "WITH recurring_tab AS(\n",
    "\tSELECT\n",
    "\t\tcustomers.customer_state AS state,\n",
    "\t\tCOUNT(*) AS num_orders,\n",
    "\t\tCASE\n",
    "\t\t\tWHEN COUNT(*)>1 THEN 1\n",
    "\t\t\tELSE 0\n",
    "\t\tEND AS is_recurring\n",
    "\tFROM orders\n",
    "\tJOIN customers ON customers.customer_id = orders.customer_id\n",
    "\tGROUP BY customers.customer_state, customers.customer_unique_id\n",
    "),\n",
    "\n",
    "pi_table AS(\n",
    "    SELECT\n",
    "        state,\n",
    "        COUNT(*) AS size,\n",
    "        SUM(is_recurring) AS recurring_size,\n",
    "        ROUND(SUM(is_recurring)::NUMERIC/COUNT(*)::NUMERIC, 4) AS pi\n",
    "    FROM recurring_tab\n",
    "    GROUP BY state\n",
    ")\n",
    "\n",
    "SELECT *\n",
    "FROM pi_table\n",
    "WHERE size*pi*(1-pi) > 15\n",
    "ORDER BY pi DESC\n",
    "'''\n",
    "\n",
    "recurring_by_state = pd.read_sql(query, engine)\n",
    "recurring_by_state.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [],
   "source": [
    "def proportion_conf_int(df, category, to_test, pi='pi', size='recurring_size', alpha=0.05, do_return=True):\n",
    "    '''Calculate the confidence interval for the difference between 2 proportions.\n",
    "    The Central Limit Theorem applies.'''\n",
    "\n",
    "    pi1 = df.loc[df[category]==to_test[0], pi].values[0]\n",
    "    pi2 = df.loc[df[category]==to_test[1], pi].values[0]\n",
    "    n1 = df.loc[df[category]==to_test[0], size].values[0]\n",
    "    n2 = df.loc[df[category]==to_test[1], size].values[0]\n",
    "\n",
    "    std_error = np.sqrt((pi1*(1-pi1) / n1) + (pi2*(1-pi2) / n2))\n",
    "    z_critical = stats.norm.ppf(1 - alpha / 2)\n",
    "    margin_of_error = z_critical * std_error\n",
    "\n",
    "    lower_bound = (pi1 - pi2) - margin_of_error\n",
    "    upper_bound = (pi1 - pi2) + margin_of_error\n",
    "\n",
    "    if do_return:\n",
    "        return lower_bound, upper_bound\n",
    "    else:\n",
    "        print(f'{to_test[0]} vs. {to_test[1]}')\n",
    "        print(f'Upper bound: {upper_bound} —— Lower bound: {lower_bound}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "conf_intervals = {'RJ_vs':[], 'up_bound':[], 'lo_bound':[]}\n",
    "prev = 'RJ'\n",
    "\n",
    "for state in recurring_by_state['state'].values:\n",
    "    if state != prev:\n",
    "        conf_intervals['RJ_vs'].append(state)\n",
    "        curr = state\n",
    "        lo, hi = proportion_conf_int(recurring_by_state, 'state', (prev, curr))\n",
    "        conf_intervals['lo_bound'].append(lo)\n",
    "        conf_intervals['up_bound'].append(hi)\n",
    "\n",
    "ci_proportion_value = pd.DataFrame(conf_intervals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.barh(ci_proportion_value['RJ_vs'], ci_proportion_value['up_bound']-ci_proportion_value['lo_bound'], left=ci_proportion_value['lo_bound'])\n",
    "plt.title('Conf. Int. % Recurring Customers by State')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [],
   "source": [
    "ci_proportion_value.to_csv(r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\recurring_pi_by_state.csv',\n",
    "                index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Payment method with the highest average order value"
   ]
  },
  {
   "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>avg_order_value</th>\n",
       "      <th>num_transactions</th>\n",
       "      <th>std</th>\n",
       "      <th>payment_type</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>27.104754</td>\n",
       "      <td>19783</td>\n",
       "      <td>39.922639</td>\n",
       "      <td>boleto</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>30.536987</td>\n",
       "      <td>76469</td>\n",
       "      <td>41.541676</td>\n",
       "      <td>credit_card</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>26.681889</td>\n",
       "      <td>1477</td>\n",
       "      <td>46.499899</td>\n",
       "      <td>debit_card</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>18.133068</td>\n",
       "      <td>1620</td>\n",
       "      <td>32.299097</td>\n",
       "      <td>voucher</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   avg_order_value  num_transactions        std payment_type\n",
       "0        27.104754             19783  39.922639       boleto\n",
       "1        30.536987             76469  41.541676  credit_card\n",
       "2        26.681889              1477  46.499899   debit_card\n",
       "3        18.133068              1620  32.299097      voucher"
      ]
     },
     "execution_count": 46,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = '''\n",
    "SELECT\n",
    "    AVG(payment_value::NUMERIC) AS avg_order_value,\n",
    "\tCOUNT(*) AS num_transactions,\n",
    "\tSTDDEV(payment_value::NUMERIC) AS std,\n",
    "\tpayment_type\n",
    "FROM order_payments\n",
    "WHERE payment_value > '0€'\n",
    "    AND payment_sequential = 1\n",
    "GROUP BY payment_type\n",
    "'''\n",
    "\n",
    "pay_types = pd.read_sql(query, engine)\n",
    "pay_types"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [],
   "source": [
    "def mean_conf_int(df, category, to_test, mean, std, size, alpha=0.05, do_return=True):\n",
    "    '''Calculate the confidence interval for the difference between 2 means.\n",
    "    The Central Limit Theorem applies.'''\n",
    "\n",
    "    mean1 = df.loc[df[category]==to_test[0], mean].values[0]\n",
    "    mean2 = df.loc[df[category]==to_test[1], mean].values[0]\n",
    "    std1 = df.loc[df[category]==to_test[0], std].values[0]\n",
    "    std2 = df.loc[df[category]==to_test[1], std].values[0]\n",
    "    n1 = df.loc[df[category]==to_test[0], size].values[0]\n",
    "    n2 = df.loc[df[category]==to_test[1], size].values[0]\n",
    "\n",
    "    std_error = np.sqrt((std1**2 / n1) + (std2**2 / n2))\n",
    "    z_critical = stats.norm.ppf(1 - alpha / 2)\n",
    "    margin_of_error = z_critical * std_error\n",
    "\n",
    "    lower_bound = (mean1 - mean2) - margin_of_error\n",
    "    upper_bound = (mean1 - mean2) + margin_of_error\n",
    "\n",
    "    if do_return:\n",
    "        return lower_bound, upper_bound\n",
    "    else:\n",
    "        print(f'{to_test[0]} vs. {to_test[1]}')\n",
    "        print(f'Upper bound: {upper_bound} —— Lower bound: {lower_bound}')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 55,
   "metadata": {},
   "outputs": [],
   "source": [
    "conf_intervals = {'CC_vs':[], 'up_bound':[], 'lo_bound':[]}\n",
    "prev = 'credit_card'\n",
    "\n",
    "for type in pay_types['payment_type'].values:\n",
    "    if type != prev:\n",
    "        lo, hi = mean_conf_int(pay_types, 'payment_type', (prev, type),\n",
    "                      'avg_order_value', 'std', 'num_transactions')\n",
    "        conf_intervals['CC_vs'].append(type)\n",
    "        conf_intervals['lo_bound'].append(lo)\n",
    "        conf_intervals['up_bound'].append(hi)\n",
    "\n",
    "ci_paym_type_value = pd.DataFrame(conf_intervals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 58,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.barh(ci_paym_type_value['CC_vs'], ci_paym_type_value['up_bound']-ci_paym_type_value['lo_bound'], left=ci_paym_type_value['lo_bound'])\n",
    "plt.title('Conf. Int. Order Value by Payment Type')\n",
    "plt.xlim(-1,15)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 59,
   "metadata": {},
   "outputs": [],
   "source": [
    "ci_paym_type_value.to_csv((r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\order_value_vs_payment_types.csv'), index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Findings:\n",
    "\n",
    "Customers that use credit card spend around 3€ more than customers that use boleto... and 11€ more than customers who use voucher."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Satisfaction vs delivery delay"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "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>delivery_status</th>\n",
       "      <th>std</th>\n",
       "      <th>frequency</th>\n",
       "      <th>average_review_score</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>on_time</td>\n",
       "      <td>1.222938</td>\n",
       "      <td>2718</td>\n",
       "      <td>4.11</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>early</td>\n",
       "      <td>1.144371</td>\n",
       "      <td>86527</td>\n",
       "      <td>4.30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>delay</td>\n",
       "      <td>1.570843</td>\n",
       "      <td>6362</td>\n",
       "      <td>2.27</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  delivery_status       std  frequency  average_review_score\n",
       "0         on_time  1.222938       2718                  4.11\n",
       "1           early  1.144371      86527                  4.30\n",
       "2           delay  1.570843       6362                  2.27"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query='''\n",
    "WITH review_vs_delivery AS(\n",
    "\tSELECT\n",
    "\t\tEXTRACT(DAY FROM orders.order_estimated_delivery_date - orders.order_delivered_customer_date) AS days_in_advance,\n",
    "\t\treviews.review_score\n",
    "\tFROM orders\n",
    "\tJOIN order_reviews AS reviews ON reviews.order_id = orders.order_id\n",
    "\tWHERE (orders.order_estimated_delivery_date - orders.order_delivered_customer_date) IS NOT NULL\n",
    ")\n",
    "\n",
    "SELECT\n",
    "    CASE\n",
    "        WHEN days_in_advance > 0 THEN 'early'\n",
    "        WHEN days_in_advance = 0 THEN 'on_time'\n",
    "        ELSE 'delay'\n",
    "    END AS delivery_status,\n",
    "\tSTDDEV(review_score) as std,\n",
    "\tCOUNT(*) AS frequency,\n",
    "    ROUND(AVG(review_score),2) AS average_review_score\n",
    "FROM \n",
    "    review_vs_delivery\n",
    "GROUP BY \n",
    "    CASE\n",
    "        WHEN days_in_advance > 0 THEN 'early'\n",
    "        WHEN days_in_advance = 0 THEN 'on_time'\n",
    "        ELSE 'delay'\n",
    "    END\n",
    "'''\n",
    "\n",
    "delivery_satisf = pd.read_sql(query, engine)\n",
    "delivery_satisf"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "delivery_satisf.to_csv(r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\bar_satisfac_vs_delay.csv',\n",
    "                 index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "conf_intervals = {'on_time_vs':[], 'up_bound':[], 'lo_bound':[]}\n",
    "prev = 'on_time'\n",
    "\n",
    "for type in delivery_satisf['delivery_status'].values:\n",
    "    if type != prev:\n",
    "        lo, hi = mean_conf_int(delivery_satisf, 'delivery_status', (prev, type),\n",
    "                      'average_review_score', 'std', 'frequency')\n",
    "        conf_intervals['on_time_vs'].append(type)\n",
    "        conf_intervals['lo_bound'].append(lo)\n",
    "        conf_intervals['up_bound'].append(hi)\n",
    "\n",
    "ci_delivery_satisf = pd.DataFrame(conf_intervals)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 70,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.barh(ci_delivery_satisf['on_time_vs'], ci_delivery_satisf['up_bound']-ci_delivery_satisf['lo_bound'], left=ci_delivery_satisf['lo_bound'])\n",
    "plt.title('Conf. Int. Delivery vs Satisfaction')\n",
    "plt.xlim(-1,2.5)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 71,
   "metadata": {},
   "outputs": [],
   "source": [
    "ci_delivery_satisf.to_csv(r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\satisfac_vs_delay.csv',\n",
    "                 index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Satisfaction Trend vs. Delivery Time"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 73,
   "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>bin</th>\n",
       "      <th>high_bound</th>\n",
       "      <th>low_bound</th>\n",
       "      <th>avg_stars</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>-21.0</td>\n",
       "      <td>-188.0</td>\n",
       "      <td>1.768725</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>1</td>\n",
       "      <td>-19.0</td>\n",
       "      <td>-20.0</td>\n",
       "      <td>1.587500</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2</td>\n",
       "      <td>-17.0</td>\n",
       "      <td>-18.0</td>\n",
       "      <td>1.672840</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>3</td>\n",
       "      <td>-15.0</td>\n",
       "      <td>-16.0</td>\n",
       "      <td>1.696000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>4</td>\n",
       "      <td>-13.0</td>\n",
       "      <td>-14.0</td>\n",
       "      <td>1.654494</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   bin  high_bound  low_bound  avg_stars\n",
       "0    0       -21.0     -188.0   1.768725\n",
       "1    1       -19.0      -20.0   1.587500\n",
       "2    2       -17.0      -18.0   1.672840\n",
       "3    3       -15.0      -16.0   1.696000\n",
       "4    4       -13.0      -14.0   1.654494"
      ]
     },
     "execution_count": 73,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = '''\n",
    "WITH review_vs_delivery AS(\n",
    "\tSELECT\n",
    "\t\tEXTRACT(DAY FROM orders.order_estimated_delivery_date - orders.order_delivered_customer_date) AS days_in_advance,\n",
    "\t\treviews.review_score\n",
    "\tFROM orders\n",
    "\tJOIN order_reviews AS reviews ON reviews.order_id = orders.order_id\n",
    "\tWHERE (orders.order_estimated_delivery_date - orders.order_delivered_customer_date) IS NOT NULL\n",
    "),\n",
    "\n",
    "reviews_delivery_bin AS(\n",
    "\tSELECT\n",
    "\t\t*,\n",
    "\t\twidth_bucket(days_in_advance, -20, 20, 20) AS bin\n",
    "\tFROM review_vs_delivery\n",
    ")\n",
    "\n",
    "SELECT\n",
    "\tbin,\n",
    "\tMAX(days_in_advance) AS high_bound,\n",
    "\tMIN(days_in_advance) AS low_bound,\n",
    "\tAVG(review_score) AS avg_stars\n",
    "FROM reviews_delivery_bin\n",
    "GROUP BY bin\n",
    "ORDER BY bin\n",
    "'''\n",
    "\n",
    "satisf_delivery_trend = pd.read_sql(query, engine)\n",
    "satisf_delivery_trend.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 76,
   "metadata": {},
   "outputs": [],
   "source": [
    "satisf_delivery_trend.loc[0, 'low_bound'] = -22\n",
    "satisf_delivery_trend.loc[21, 'high_bound'] = 21"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 77,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAiMAAAGxCAYAAACwbLZkAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjguMCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy81sbWrAAAACXBIWXMAAA9hAAAPYQGoP6dpAABNLUlEQVR4nO3deVyU5f4//tfALAzbsG8Cghug5Iqmllukpkl5suW0mmZlaWbm75ie+lmfFso651hZmi2aWmodtGO5pJZgixooueOSIsiOyg4DM3N9/4AZHQFlcOCeYV7Px2MeOvdcM/d7uIF5cd3Xdd0yIYQAERERkUScpC6AiIiIHBvDCBEREUmKYYSIiIgkxTBCREREkmIYISIiIkkxjBAREZGkGEaIiIhIUgwjREREJCmGESIiIpIUwwi1yr59+/C3v/0N4eHhUKlUCAwMxJAhQ/Diiy+26vW2bNmCV199tcnHIiIi8Pjjj1v8muvXr0evXr2gVqshk8nw559/tqq25uTm5uLVV19t8nVfffVVyGQyq+7veoz7vN5t5MiR7VpXU2QyWbPHW2pXfq2cnZ3h7e2NPn364Omnn8bevXtv6LWv/l7OzMyETCbDypUrb6zodhYREdGi77WVK1dK8rNA9kcudQFkfzZv3oy77roLI0eOxKJFixAcHIy8vDykpaVh3bp1+Ne//mXxa27ZsgUfffRRkx9QGzduhKenp0WvV1RUhEcffRR33HEHPv74Y6hUKvTo0cPiuq4lNzcXr732GiIiItC3b1+zx6ZNm4Y77rjDqvu7nqv3mZeXh3vuuQfPPfccHnroIdN2S7+Wjujee+/Fiy++CCEEysrKcOTIEaxatQrLly/HrFmz8P7771tlP8HBwdizZw+6du1qlddrLxs3boRWqzXd/+yzz/D5559j27Zt0Gg0pu1du3aFVqtt958Fsj8MI2SxRYsWITIyEj/++CPk8svfQn//+9+xaNEiq++vX79+Fj/n5MmTqKurwyOPPIIRI0ZYvabrCQ0NRWhoqKT7zMzMBACEh4dj8ODBzT6vrq4OMpnM7Fg6usDAQLOv2dixYzF79mw89dRT+OCDDxAdHY1nnnnmhvejUqmueWzaSnV1NVxcXFrdY3H1z+S2bdsAAAMGDICfn1+j9u39s0D2h6dpyGIXLlyAn59fkx9eTk7m31Lr16/HmDFjEBwcDLVajZiYGLz00kuorKw0tXn88cfx0UcfATDvIjd+mF7dtW0wGPDGG28gKioKarUaXl5e6N27t+mv1ccffxy33norAOCBBx4wOzWRlpaGv//974iIiIBarUZERAQefPBBnDt3rtF7ycnJwVNPPYWwsDAolUqEhITg3nvvRUFBAZKTkzFw4EAAwJQpU0w1G3t2muqaNhgMWLRoEaKjo6FSqRAQEIDHHnsM58+fN2s3cuRIxMbGIjU1FcOGDYOrqyu6dOmCt99+GwaD4VqH5rqSk5Mhk8mwevVqvPjii+jUqRNUKhVOnz4NANi5cyfi4+Ph6ekJV1dX3HLLLfjpp5/MXsP43o4ePYoHH3wQGo0GgYGBmDp1KkpLS83alpWV4cknn4Svry/c3d1xxx134OTJk9ets6ioCEqlEq+88kqjxzIyMiCTyfDBBx8AAKqqqjB37lxERkbCxcUFPj4+iIuLw9q1a1v7ZWqSs7MzlixZAj8/P7z77rtmj5WVlZlqUCqV6NSpE2bPnm32fd6Uq0/TfPfdd5DJZI2+5gCwdOlSyGQyHDp0yLQtLS0Nd911F3x8fODi4oJ+/frhm2++MXveypUrIZPJsH37dkydOhX+/v5wdXXFr7/+CplM1uTXadWqVZDJZEhNTW3pl6dZTf0sREREYMKECfjhhx/Qr18/0++GH374wVRzTEwM3NzcMGjQIKSlpTV63Za8d7IfDCNksSFDhmDfvn2YNWsW9u3bh7q6umbbnjp1CuPHjzd14c6ePRvffPMNEhISTG1eeeUV3HvvvQCAPXv2mG7BwcFNvuaiRYvw6quv4sEHH8TmzZuxfv16PPHEEygpKTG9njHcvPXWW9izZw8+/vhjAPW//KOiorB48WL8+OOPeOedd5CXl4eBAweiuLjYtI+cnBwMHDgQGzduxJw5c7B161YsXrwYGo0Gly5dQv/+/bFixQoAwMsvv2yqedq0ac1+LZ555hnMmzcPo0ePxqZNm/D6669j27ZtGDp0qNm+ASA/Px8PP/wwHnnkEWzatAnjxo3D/PnzsWbNmmZf3xLz589HVlYWli1bhu+//x4BAQFYs2YNxowZA09PT3z55Zf45ptv4OPjg7Fjxzb54Thp0iT06NEDSUlJeOmll/D111/jhRdeMD0uhMDEiRNNwWfjxo0YPHgwxo0bd936/P39MWHCBHz55ZeNAtiKFSugVCrx8MMPAwDmzJmDpUuXYtasWdi2bRtWr16N++67DxcuXLjBr1JjarUat99+O86ePWsKkVVVVRgxYgS+/PJLzJo1C1u3bsW8efOwcuVK3HXXXbDkwugTJkxAQECA6XvrSitXrkT//v3Ru3dvAMCuXbtwyy23oKSkBMuWLcP//vc/9O3bFw888ECTY1CmTp0KhUKB1atX47///S+GDh2Kfv36mX5WrrRkyRIMHDjQFLjbwsGDBzF//nzMmzcPGzZsgEajwT333IOFCxfis88+w1tvvYWvvvoKpaWlmDBhAqqrq03PtfS9kx0QRBYqLi4Wt956qwAgAAiFQiGGDh0qEhMTRXl5ebPPMxgMoq6uTqSkpAgA4uDBg6bHZsyYIZr7duzcubOYPHmy6f6ECRNE3759r1njrl27BADx7bffXrOdTqcTFRUVws3NTbz//vum7VOnThUKhUIcO3as2eempqYKAGLFihWNHlu4cKHZ+zl+/LgAIJ599lmzdvv27RMAxIIFC0zbRowYIQCIffv2mbXt2bOnGDt27DXfz5XOnj0rAIh3333XtM34dRk+fLhZ28rKSuHj4yMSEhLMtuv1etGnTx8xaNCgRu9t0aJFZm2fffZZ4eLiIgwGgxBCiK1btwoAZl9XIYR48803BQCxcOHCa9a/adMmAUBs377dtE2n04mQkBAxadIk07bY2FgxceLEa76WJQCIGTNmNPv4vHnzzI5PYmKicHJyEqmpqWbt/vvf/woAYsuWLaZtV38vG4/Rld9Dc+bMEWq1WpSUlJi2HTt2TAAQH374oWlbdHS06Nevn6irqzPb74QJE0RwcLDQ6/VCCCFWrFghAIjHHnus0XsxPpaenm7a9scffwgA4ssvv2z2a3A14/dEUVFRs49dqXPnzkKtVovz58+btv35558CgAgODhaVlZWm7d99950AIDZt2mTxeyf7wZ4Rspivry9++eUXpKam4u2338bdd9+NkydPYv78+bjpppvM/so/c+YMHnroIQQFBcHZ2RkKhcI0huP48eOt2v+gQYNw8OBBPPvss/jxxx9RVlbW4udWVFRg3rx56NatG+RyOeRyOdzd3VFZWWlWz9atWzFq1CjExMS0qsar7dq1CwAazQoaNGgQYmJiGvU8BAUFYdCgQWbbevfu3eTppNaYNGmS2f3ff/8dFy9exOTJk6HT6Uw3g8GAO+64A6mpqY1OOdx1112N6qupqUFhYSGAy+/Z2INhdOVg2msZN24cgoKCzHoJfvzxR+Tm5mLq1KmmbYMGDcLWrVvx0ksvITk52ewv6LYgrurp+OGHHxAbG4u+ffuafe3Gjh0LmUyG5ORki15/6tSpqK6uxvr1603bVqxYAZVKZfranT59GhkZGaav7ZX7HT9+PPLy8nDixAmz1736mAPAgw8+iICAALPekQ8//BD+/v544IEHLKrbUn379kWnTp1M940/ayNHjoSrq2uj7cbv/da8d7J9DCPUanFxcZg3bx6+/fZb5Obm4oUXXkBmZqZpEGtFRQWGDRuGffv24Y033kBycjJSU1OxYcMGAGj1h8b8+fPx3nvvYe/evRg3bhx8fX0RHx/f5Hnlqz300ENYsmQJpk2bhh9//BF//PEHUlNT4e/vb1ZPUVGRVQfdGU8ZNHXqKSQkpNEpBV9f30btVCqV1T5or66joKAAQP0sEoVCYXZ75513IITAxYsXr1mjSqUCcPm4XrhwAXK5vFG7oKCgFtUol8vx6KOPYuPGjaZTcCtXrkRwcDDGjh1ravfBBx9g3rx5+O677zBq1Cj4+Phg4sSJOHXqVIv2Yynjh2JISAiA+q/doUOHGn3dPDw8IIRodAruenr16oWBAweaQpher8eaNWtw9913w8fHx7RPAJg7d26j/T777LMA0Gi/TX3vqVQqPP300/j6669RUlKCoqIifPPNN5g2bZrpeLYV43sxUiqV19xeU1MDoHXvnWwfh8+TVSgUCixcuBD/+c9/cOTIEQDAzz//jNzcXCQnJ5vNaDF+sLSWXC7HnDlzMGfOHJSUlGDnzp1YsGABxo4di+zsbLO/qq5UWlqKH374AQsXLsRLL71k2q7Vaht90Pr7+zcaWHojjB/IeXl5jUJObm5ukzMQ2tLVAwqN+//www+bnd0RGBho0T58fX2h0+lw4cIFs0CSn5/f4teYMmUK3n33Xaxbtw4PPPAANm3ahNmzZ8PZ2dnUxs3NDa+99hpee+01FBQUmHpJEhISkJGRYVHN11NdXY2dO3eia9eupuPo5+cHtVqNL774osnntObYTpkyBc8++yyOHz+OM2fOIC8vD1OmTGn0mvPnz8c999zT5GtERUWZ3W9u5swzzzyDt99+G1988QVqamqg0+kwffp0i2tuL61572T7GEbIYnl5eU3+lWU8zWH8i9H4y+/qv7A++eSTRs+98q9qtVrd4lq8vLxw7733IicnB7Nnz0ZmZiZ69uzZZFuZTAYhRKN6PvvsM+j1erNt48aNw+rVq3HixIlmf7Fd3RNwLbfddhsAYM2aNWaDAlNTU3H8+HH885//vO5rtKVbbrkFXl5eOHbsGGbOnGmV1xw1ahQWLVqEr776CrNmzTJt//rrr1v8GjExMbj55puxYsUK6PV6aLVasw/lqwUGBuLxxx/HwYMHsXjxYlRVVTUbTi2l1+sxc+ZMXLhwAYmJiabtEyZMwFtvvQVfX19ERkZaZV8PPvgg5syZg5UrV+LMmTPo1KkTxowZY3o8KioK3bt3x8GDB/HWW2/d0L6Cg4Nx33334eOPP0ZtbS0SEhIQHh5+o2+hzVjzvZPtYBghi40dOxahoaFISEhAdHQ0DAYD/vzzT/zrX/+Cu7s7nn/+eQDA0KFD4e3tjenTp2PhwoVQKBT46quvcPDgwUavedNNNwEA3nnnHYwbNw7Ozs7o3bu3qYv2SgkJCYiNjUVcXBz8/f1x7tw5LF68GJ07d0b37t2brdvT0xPDhw/Hu+++Cz8/P0RERCAlJQWff/45vLy8zNr+3//9H7Zu3Yrhw4djwYIFuOmmm1BSUoJt27Zhzpw5iI6ORteuXaFWq/HVV18hJiYG7u7uCAkJMYWxK0VFReGpp57Chx9+CCcnJ4wbNw6ZmZl45ZVXEBYWZjYLRQru7u748MMPMXnyZFy8eBH33nsvAgICUFRUhIMHD6KoqAhLly616DXHjBmD4cOH4x//+AcqKysRFxeH3377DatXr7bodaZOnYqnn34aubm5GDp0aKNwePPNN2PChAno3bs3vL29cfz4caxevRpDhgwxBZFVq1Zh6tSp+OKLL/DYY49dd58FBQXYu3cvhBAoLy83LXp28OBBvPDCC3jyySdNbWfPno2kpCQMHz4cL7zwAnr37g2DwYCsrCxs374dL774Im6++WaL3rOXlxf+9re/YeXKlSgpKcHcuXMbTZv/5JNPMG7cOIwdOxaPP/44OnXqhIsXL+L48eM4cOAAvv322xbv7/nnnzfV2NRMHltjzfdONkLS4bNkl9avXy8eeugh0b17d+Hu7i4UCoUIDw8Xjz76aKPZJ7///rsYMmSIcHV1Ff7+/mLatGniwIEDjWYQaLVaMW3aNOHv7y9kMpkAIM6ePSuEaDwD4V//+pcYOnSo8PPzE0qlUoSHh4snnnhCZGZmmto0N5vm/PnzYtKkScLb21t4eHiIO+64Qxw5cqTRPoQQIjs7W0ydOlUEBQUJhUIhQkJCxP333y8KCgpMbdauXSuio6OFQqEwmyHS1AwCvV4v3nnnHdGjRw+hUCiEn5+feOSRR0R2drZZuxEjRohevXo1+rpPnjxZdO7cualD0qRrzaZpbpZRSkqKuPPOO4WPj49QKBSiU6dO4s477zRr39zMCePMDONxE0KIkpISMXXqVOHl5SVcXV3F6NGjRUZGRotm0xiVlpYKtVotAIhPP/200eMvvfSSiIuLE97e3kKlUokuXbqIF154QRQXFzeqramZT1dDwywxAMLJyUl4enqKm266STz11FNiz549TT6noqJCvPzyyyIqKkoolUqh0WjETTfdJF544QWRn59vateS2TRG27dvN9Vx8uTJJvd78OBBcf/994uAgAChUChEUFCQuO2228SyZcsavferZ/tcLSIiQsTExFyzTXNaM5vmzjvvbNQWTcxkaur7WIiWvXeyHzIhLJgET0REHc6hQ4fQp08ffPTRR6ZBoETtiWGEiMhB/fXXXzh37hwWLFiArKwsnD592mpjbIgswam9REQO6vXXX8fo0aNRUVGBb7/9lkGEJMOeESIiIpLUDfWMJCYmQiaTYfbs2c22MV6Y6+qbtef/ExERkX1q9dTe1NRULF++3HTRpus5ceIEPD09Tff9/f1bu2siIiLqQFrVM1JRUYGHH34Yn376Kby9vVv0nICAAAQFBZluV66gSERERI6rVT0jM2bMwJ133onbb78db7zxRoue069fP9TU1KBnz554+eWXMWrUqGbbarVaaLVa032DwYCLFy/C19e32SWNiYiIyLaIhoUDQ0JCGi3cdyWLw8i6detw4MABpKamtqh9cHAwli9fjgEDBkCr1WL16tWIj49HcnIyhg8f3uRzEhMT8dprr1laGhEREdmg7Ozsa1581KLZNNnZ2YiLi8P27dvRp08fAPWXe+7bty8WL17c4qISEhIgk8mwadOmJh+/umektLQU4eHhyM7ONht3QkRERLarrKwMYWFhKCkpgUajabadRT0j+/fvR2FhIQYMGGDaptfrsXv3bixZsgRarbZFY0EGDx6MNWvWNPu4SqVq8vLVnp6eDCNERER25npDLCwKI/Hx8Th8+LDZtilTpiA6Ohrz5s1r8aDU9PT0Jq/6SkRERI7HojDi4eGB2NhYs21ubm7w9fU1bZ8/fz5ycnKwatUqAMDixYsRERGBXr16oba2FmvWrEFSUhKSkpKs9BaIiIjInrV6nZHm5OXlISsry3S/trYWc+fORU5ODtRqNXr16oXNmzdj/Pjx1t41ERER2SG7WA6+rKwMGo0GpaWlHDNCRERkJ1r6+c0L5REREZGkGEaIiIhIUgwjREREJCmGESIiIpIUwwgRERFJimGEiIiIJMUwQkRERJJiGCEiIiJJWX0FViIiIrKeCq0OfxVW4HRhBTIvVCLSzw3jbwqGi6Jl14OzBwwjREREEhNC4EJlLU43hI7ThRX4q6j+37zSmkbt39h8HH8fGIZHBndGiJdagoqti8vBExGRQ6vVGSCTAQrnth+5YDAI5JRU43RRham343RhBU4XVaCkqq7Z5/m5q9AtwA3hPq749VQxchsCipMMGNMzCJOHRmBwFx/IZLI2fw+WaOnnN3tGiIjI7ugNAhVaHSq1OlRodSivufz/ihodyq/xWIXW/FarMwAAFM4yqBXOcFXK4ap0hlrp3PCvHK6K+vtqpTNcFVdsv6Kdq9IZaoXcdF8I4EzR5bBxurACZ4oqUV2nb/I9yWRAqLca3fzd0S3gipu/BzSuClM7nd6AnccL8OXv57DnzAVsO5qPbUfzERXogceGdsbf+nWCq9K+Pt7ZM0JERDZHq9Mj60IV/iqqxNniSpwpqsCZ4kqcv1SF8hodqmqb/kC3BwpnGSL93BqChju6NoSOLn7uUCstGwdyIr8cq/ZkYsOBHFPI8XCR4/64MDw6uDMi/Nza4i20WEs/vxlGiIhIEkII5JfV4ExRJc40BI764FEfOgwt+HRSOMvg4aKAm8oZ7ioFPFTy+v+7KOCuksPDRQ43pRzuLvKGx+r/b3pMJYe7Ug4BgapaParr9Kiu1aOqVo+qWt3l/9fpUV1bH4IuP65Hdd3lNtV1+ise18EggAhfV1PYMPZ4hPu4Qm7lU0Kl1XX47/7zWL0nE5kXqgDU97SM7OGPyUMjMLy7P5yc2v8UDsMIERFdV4VWhyM5pZABUCmcoXR2glLuBFXDrf7/zlDKneDcyg+z8po6nLmqh8N4v7lTFgDgrpKji78buvi5IdLPHV383dDZ1xVeaiXcXepDh0recWaUWIPBIJByqghf/p6J5BNFpu2Rfm54dHBn3BsXCk8XxTVewboYRoiIqEl5pdXYebwQO44VYO9fF1CrN7Toec5OMiidnaBSOJmFFmVDWFFd9diFylqcLa5EUbm22deUO8kQ7uOKSD+3+uDh7276v7+7yuYGZNqTs8WVWL3nHL5Ny0a5VgcAcFU6457+nTB5SAS6B3q0eQ0MI0REBKD+dMixvDLsPFaIHcfzcSSnzOzxEI0LXJTO0NYZUKs3oFZngFanR63O0KJTJS3h76FCpJ8buvq71YeNhp6OMB/XdpnF4sgqtTpsTM/Bqj2ZOFlQYdo+tKsvJg+NQHx0gNVPGxkxjBARObBanQH7zl7AzmMF2Hm8EDkl1abHZDKgf7g3bo8JxOieAejq795sD4ROb4BWVx9QavWGhsCih1Z3xfYr/6/Xm+57uMjRxc8dkf5u7XpqgJomhMCeMxfw5e+Z2HGswBQ0O3mp8cjgznhgYBh83JRW3Sen9hIROZjSqjokn6w//ZJyosjUNQ8ALgonDOvuj9ExgRgVHQB/D1WLXlPu7AS5sxPcWtacbJhMJsPQrn4Y2tUPOSXVWLP3HNb9kYWckmq8sy0DAgLPjuwmTW3sGSEisl/ZF6uw41gBdh4vwB9nL0J3xXkVP3cVbo8JwO0xgbilm5/F00ap46up0+P7g7lYl5qNTx+LY88IERFdn8EgcCintOH0SwEy8svNHu8e4I7RPQNxe89A9A31kmQ6J9kPF4Uz7osLw31xYZLWwTBCRGSjaur0pnU3zhTVr+K5568LKLxidoqTDBgY4VMfQGICJV/kiqg1GEaIiCRktvBXUQX+KqrEX0X1y4bnllajqRPpbkpnjIjyx+iegRjZIwDeVu5aJ2pvDCNERO2gqlZnttLomYbQcba48ppLm3u6yNHFv34abFd/d8R20mBwFx8u9kUdCsMIEZGVlVTV4vtDeThVUG4KHU1dBt7IuWHhry5XLPzVtSGA+LopufAXdXgMI0REVlJdq8cXv53FspS/UF6ja/S4t6uivpfDzxg46v8N93GFUs6Fv8hxMYwQEd2gOr0B61Oz8cFPp0yDS6MCPTAy2h9dG1Ya7eLvbvVpk0QdBcMIEVErGQwCW47k4V/bT+JscSUAINRbjRfH9MDdfTpxWi1RCzGMEBG1wq+nivHOtgwczikFAPi6KfHcbd3w4M3hHFxKZCGGESIiCxw+X4p3tmXg19PFAOqn2T45vAumDesCdxV/pRK1Bn9yiIha4GxxJd7bfgKbD+UBABTOMjx8c2fMvK0b/Nx54RaiG8EwQkR0DYVlNVj80ymsT82G3iAgkwET+3bCnNE9EObjKnV5RB0CwwgRURNKq+vwScpf+OK3s6ipMwAARkX54x93RCMmmBfsJLImhhEioivU1Omxak8mPtr1F0qr6wAA/cO9MO+OaNzcxVfi6og6JoYRIiIAOr0BSQfOY/HOU6bVUrsHuOP/GxuF0T0DuQoqURtiGCEihyaEwI9HC/De9hM4XVgBAAjRuGD26B6Y1D8UzlwrhKjNMYwQUYdXU6dHTkk1ci5VN/o362IV8svqe0K8XBWYOaobHhncGS4KrhVC1F4YRojI7pVW110RMKrq/70idBRX1F7z+WqFM564NRJPjegCTxdFO1VNREYMI0RkF3R6A5JPFOFMcYUpZJxv+Lepi9JdzVXpjE5eanTyVpv9G+qtRrcAD2jUDCFEUmEYISK7sPaPLLzyv6PNPu7tqrgcNLxczcJGJy81vFwVHIRKZKMYRojILmw9kg8AGBTpg7jO3ghp6N0IbfjXVclfZ0T2ij+9RGTzymrq8MfZiwCAd+/tjc6+bhJXRETW5CR1AURE1/PrqWLoDAJd/d0YRIg6IIYRIrJ5Px0vBADcFh0gcSVE1BYYRojIpukNAsknjGEkUOJqiKgtMIwQkU07eL4EFypr4eEiR1yEt9TlEFEbYBghIpu2K6O+V2RED38onPkri6gj4k82Edk0jhch6vgYRojIZuWVVuNYXhlkMmBkFMMIUUfFMEJENmtXRhEAoH+4N3zclBJXQ0Rt5YbCSGJiImQyGWbPnn3NdikpKRgwYABcXFzQpUsXLFu27EZ2S0QO4ueMAgA8RUPU0bU6jKSmpmL58uXo3bv3NdudPXsW48ePx7Bhw5Ceno4FCxZg1qxZSEpKau2uicgB1NTp8evpYgAMI0QdXavCSEVFBR5++GF8+umn8Pa+9lS7ZcuWITw8HIsXL0ZMTAymTZuGqVOn4r333mtVwUTkGPb8dQE1dQaEaFwQHeQhdTlE1IZaFUZmzJiBO++8E7fffvt12+7Zswdjxowx2zZ27FikpaWhrq6uyedotVqUlZWZ3YjIsfzcMKX3tpgAXm2XqIOzOIysW7cOBw4cQGJiYova5+fnIzDQfNXEwMBA6HQ6FBcXN/mcxMREaDQa0y0sLMzSMonIjgkhLocRnqIh6vAsCiPZ2dl4/vnnsWbNGri4uLT4eVf/VSOEaHK70fz581FaWmq6ZWdnW1ImEdm5EwXlyCmphovCCUO7+kldDhG1Mbkljffv34/CwkIMGDDAtE2v12P37t1YsmQJtFotnJ2dzZ4TFBSE/Px8s22FhYWQy+Xw9fVtcj8qlQoqlcqS0oioAzEudHZLVz+4KJyv05qI7J1FYSQ+Ph6HDx822zZlyhRER0dj3rx5jYIIAAwZMgTff/+92bbt27cjLi4OCoWiFSUTUUe364rxIkTU8VkURjw8PBAbG2u2zc3NDb6+vqbt8+fPR05ODlatWgUAmD59OpYsWYI5c+bgySefxJ49e/D5559j7dq1VnoLRNSRXKysxYGsSwCAUVx1lcghWH0F1ry8PGRlZZnuR0ZGYsuWLUhOTkbfvn3x+uuv44MPPsCkSZOsvWsi6gBSThbCIICYYE+EeKmlLoeI2oFFPSNNSU5ONru/cuXKRm1GjBiBAwcO3OiuiMgBGMeLxHMWDZHD4LVpiMhm1OkN2H2y/no0oxhGiBwGwwgR2Yz95y6hrEYHHzcl+oZ5SV0OEbUThhEishnGhc5GRvnD2YmrrhI5CoYRIrIZPx2vv0pvfHTgdVoSUUfCMEJENuHchUr8VVQJuZMMw3pw1VUiR8IwQkQ2wXiKZmCEDzxduCAikSNhGCEim2AMI/FcdZXI4TCMEJHkKrQ67DtzEQCv0kvkiBhGiEhyv54qRq3egAhfV3Txd5e6HCJqZwwjRCS5nzPqZ9Hcxlk0RA6JYYSIJGUwCPycUb/qKseLEDkmhhEiktSR3FIUV2jhrpJjYISP1OUQkQQYRohIUsYL4w3r7gelnL+SiBwRf/KJSFLGKb2cRUPkuBhGiEgyhWU1OJxTCpkMGBnFMELkqBhGiEgyu07U94r0DvWCv4dK4mqISCoMI0QkGeN4kXieoiFyaAwjRCSJmjo9fj1dDIDjRYgcHcMIEUli39mLqKrVI9BThV4hnlKXQ0QSYhghIknsumIWjUwmk7gaIpISwwgRtTshBH7iEvBE1IBhhIja3enCCmRfrIZS7oRbuvlKXQ4RSYxhhIjanXGhsyFdfOGqlEtcDRFJjWGEiNrdTw1hhBfGIyKAYYSI2llJVS32n7sEABjFVVeJCAwjRNTOUk4WQW8QiAr0QJiPq9TlEJENYBghonZlnNI7igudEVEDhhEiajc6vQHJJ4sAcLwIEV3GMEJE7SY9uwQlVXXwclWgX5iX1OUQkY1gGCGidmO8MN6IHv6QO/PXDxHV428DImo3Vy4BT0RkxDBCRO0i+2IVThSUw9lJhhE9/KUuh4hsCMMIEbWLXSfqe0UGdPaGl6tS4mqIyJYwjBBRuzCOF+EpGiK6GsMIEbW5qlod9py5AACIZxghoqswjBBRm/vt9AXU6gwI81GjW4C71OUQkY1hGCGiNvdzRgEAID46EDKZTOJqiMjWMIwQUZsSQuBnLgFPRNfAMEJEbepobhkKyrRwVTrj5kgfqcshIhvEMEJEbcrYK3JrNz+4KJwlroaIbBHDCBG1qZ+46ioRXQfDCBG1maJyLQ6dLwHA8SJE1DyGESJqM8knCiEEcFMnDQI9XaQuh4hsFMMIEbWZn3mKhohagGGEiNpErc6AX04VA2AYIaJrYxghojaRmnkRFVod/NxVuKmTRupyiMiGMYwQUZu4fGE8fzg5cdVVImoewwgRWZ0QAj81LAF/W3SgxNUQka1jGCEiqztTXIlzF6qgcJbh1u5+UpdDRDbOojCydOlS9O7dG56envD09MSQIUOwdevWZtsnJydDJpM1umVkZNxw4URku3Y1zKIZ3MUX7iq5xNUQka2z6LdEaGgo3n77bXTr1g0A8OWXX+Luu+9Geno6evXq1ezzTpw4AU9PT9N9f3//VpZLRPbg8ngRzqIhouuzKIwkJCSY3X/zzTexdOlS7N2795phJCAgAF5eXq0qkIjsS53egP3nLgEARvTgHx5EdH2tHjOi1+uxbt06VFZWYsiQIdds269fPwQHByM+Ph67du267mtrtVqUlZWZ3YjIPpwqqECt3gAPFzki/dykLoeI7IDFYeTw4cNwd3eHSqXC9OnTsXHjRvTs2bPJtsHBwVi+fDmSkpKwYcMGREVFIT4+Hrt3777mPhITE6HRaEy3sLAwS8skIokcyS0FAPQK8YRMxim9RHR9MiGEsOQJtbW1yMrKQklJCZKSkvDZZ58hJSWl2UBytYSEBMhkMmzatKnZNlqtFlqt1nS/rKwMYWFhKC0tNRt7QkS259VNR7Hy90xMuzUSL09o2e8FIuqYysrKoNForvv5bfEwd6VSaRrAGhcXh9TUVLz//vv45JNPWvT8wYMHY82aNddso1KpoFKpLC2NiGzAkZyGnpFO/MOBiFrmhtcZEUKY9WJcT3p6OoKDg290t0RkgwwGgeN59WO8eoVwCXgiahmLekYWLFiAcePGISwsDOXl5Vi3bh2Sk5Oxbds2AMD8+fORk5ODVatWAQAWL16MiIgI9OrVC7W1tVizZg2SkpKQlJRk/XdCRJLLvFCJylo9XBRO6MLBq0TUQhaFkYKCAjz66KPIy8uDRqNB7969sW3bNowePRoAkJeXh6ysLFP72tpazJ07Fzk5OVCr1ejVqxc2b96M8ePHW/ddEJFNOJJb3ysSHeQJuTMXeCailrF4AKsUWjoAhoiklbj1OD5JOYNHBofjjYk3SV0OEUmspZ/f/NOFiKzmaA7HixCR5RhGiMgqhBA42rDGSCzDCBFZgGGEiKwit7QGl6rqIHeSoUeQu9TlEJEdYRghIqs42rC+SLcAd6jkzhJXQ0T2hGGEiKzCOJMmthNP0RCRZRhGiMgqjl1xTRoiIkswjBCRVRxlzwgRtRLDCBHdsAsVWuSV1kAmA2KC2TNCRJZhGCGiG2bsFYn0dYO7yuLrbxKRg2MYIaIbdqRhvEhPjhcholZgGCGiG2bsGeHKq0TUGgwjRHTDjGuMxHZizwgRWY5hhIhuSHlNHTIvVAFgzwgRtQ7DCBHdkON55QCAEI0LfNyUEldDRPaIYYSIbsiRHOPgVfaKEFHrMIwQ0Q25vNgZx4sQUeswjBDRDTlqWgaePSNE1DoMI0TUajV1epwqrADAa9IQUesxjBBRq53IL4feIODjpkSwxkXqcojITjGMEFGrXV7szBMymUziaojIXjGMEFGrHeF4ESKyAoYRImq1K3tGiIhai2GEiFpFpzcgI884rZc9I0TUegwjRNQqfxVVQqszwF0lR2cfV6nLISI7xjBCRK1iXF8kJtgDTk4cvEpErccwQkStciTHOF6Ep2iI6MYwjBBRq1xeeZWDV4noxjCMEJHFDAaBY7kcvEpE1sEwQkQWy75UhXKtDkq5E7oFuEtdDhHZOYYRIrKYcX2R6CAPKJz5a4SIbgx/ixCRxY7kcLwIEVkPwwgRWczYM9KTM2mIyAoYRojIIkII00yaWPaMEJEVMIwQkUUKy7UorqiFkwyIDmIYIaIbxzBCRBYxjhfpFuAOtdJZ4mqIqCNgGCEii1y+Ui/HixCRdTCMEJFFOJOGiKyNYYSILMKeESKyNoYRImqxkqpa5JRUAwB6smeEiKyEYYSIWszYKxLu4wqNWiFxNUTUUTCMEFGL8Uq9RNQWGEaIqMWO5PBKvURkfQwjRNRixp4RjhchImtiGCGiFqnU6nCmuBIAEMuZNERkRQwjRNQiGfllEAII8FDB30MldTlE1IEwjBBRi1xeX4SnaIjIuhhGiKhFjCuvcvAqEVkbwwgRtQh7RoiorTCMENF11eoMOFlQDoDLwBOR9TGMENF1nSwoR51eQKNWINRbLXU5RNTBWBRGli5dit69e8PT0xOenp4YMmQItm7des3npKSkYMCAAXBxcUGXLl2wbNmyGyqYiNrflSuvymQyiashoo7GojASGhqKt99+G2lpaUhLS8Ntt92Gu+++G0ePHm2y/dmzZzF+/HgMGzYM6enpWLBgAWbNmoWkpCSrFE9E7YPjRYioLcktaZyQkGB2/80338TSpUuxd+9e9OrVq1H7ZcuWITw8HIsXLwYAxMTEIC0tDe+99x4mTZrU+qqJqF0ZZ9JwvAgRtYVWjxnR6/VYt24dKisrMWTIkCbb7NmzB2PGjDHbNnbsWKSlpaGurq7Z19ZqtSgrKzO7EZE09AaB43n1g1djO7FnhIisz+IwcvjwYbi7u0OlUmH69OnYuHEjevbs2WTb/Px8BAYGmm0LDAyETqdDcXFxs/tITEyERqMx3cLCwiwtk4is5GxxJarr9FArnBHp5y51OUTUAVkcRqKiovDnn39i7969eOaZZzB58mQcO3as2fZXD3YTQjS5/Urz589HaWmp6ZadnW1pmURkJcbBqzHBHnB24uBVIrI+i8aMAIBSqUS3bt0AAHFxcUhNTcX777+PTz75pFHboKAg5Ofnm20rLCyEXC6Hr69vs/tQqVRQqXjtCyJbcHnwKseLEFHbuOF1RoQQ0Gq1TT42ZMgQ7Nixw2zb9u3bERcXB4VCcaO7JqJ2cHkZeI4XIaK2YVEYWbBgAX755RdkZmbi8OHD+Oc//4nk5GQ8/PDDAOpPrzz22GOm9tOnT8e5c+cwZ84cHD9+HF988QU+//xzzJ0717rvgojahBCCPSNE1OYsOk1TUFCARx99FHl5edBoNOjduze2bduG0aNHAwDy8vKQlZVlah8ZGYktW7bghRdewEcffYSQkBB88MEHnNZLZCfOX6pGaXUd5E4ydA/k4FUiahsyYRxRasPKysqg0WhQWloKT092FRO1l21H8jF9zX70DPbElueHSV0OEdmZln5+89o0RNSsY1csA09E1FYYRoioWUcaxovEduJ4ESJqOwwjRNSso+wZIaJ2wDBCRE0qKteioEwLmQyICWYYIaK2wzBCRE0y9opE+rnBTWXx+ohERC3GMEJETeL6IkTUXhhGiKhJxp6RWI4XIaI2xjBCRE1izwgRtReGESJqpKymDucuVAHgTBoiansMI0TUyLGGXpFOXmp4uyklroaIOjqGESJqxHil3p7sFSGidsAwQkSNGHtGYjlehIjaAcMIETVyhCuvElE7YhghIjPVtXqcLqwAwGvSEFH7YBghIjMZ+WUwCMDXTYlAT5XU5RCRA2AYISIzpvVFOmkgk8kkroaIHAHDCBGZubzYGceLEFH7YBghIjNHOXiViNoZwwgRmdTpDcjILwfAab1E1H4YRojI5HRhBWp1Brir5Aj3cZW6HCJyEAwjRGRiHC/SM8QTTk4cvEpE7YNhhIhMjMvAc7wIEbUnhhEiMuEy8EQkBYYRIgIAGAwCx/KMa4ywZ4SI2g/DCBEBAM5drEKFVgel3Ald/d2lLoeIHAjDCBEBuLy+SEyQBxTO/NVARO2Hv3GICABwJMc4k4bjRYiofTGMEBGAyz0jsRwvQkTtjGGEiCCEuOKaNOwZIaL2xTBCRMgvq8HFylo4O8kQHeQhdTlE5GAYRojINF6km787XBTOEldDRI6GYYSIeKVeIpIUwwgRXR4v0onjRYio/TGMEBGO8po0RCQhhhEiB3exsha5pTUA6q/WS0TU3hhGiByccbxIZ19XeLooJK6GiBwRwwiRgzvKK/USkcQYRogc3JGG8SI8RUNEUmEYIXJwx0wrrzKMEJE0GEaIHFiFVoezFyoBcBl4IpIOwwiRAzueVwYhgEBPFfw9VFKXQ0QOimGEyIEdOHcJAAevEpG0GEaIHNhPxwsBAMO6+0lcCRE5MoYRIgd1qbIWaecuAgDiYwIlroaIHBnDCJGDSj5ZCIMAogI9EObjKnU5ROTAGEaIHNTOhlM08TEBEldCRI6OYYTIAdXqDNh9oggAT9EQkfQYRogcUGrmRZRrdfB1U6JvmJfU5RCRg2MYIXJAO48XAABGRQfA2UkmcTVE5OgsCiOJiYkYOHAgPDw8EBAQgIkTJ+LEiRPXfE5ycjJkMlmjW0ZGxg0VTkStI4QwTem9neNFiMgGWBRGUlJSMGPGDOzduxc7duyATqfDmDFjUFlZed3nnjhxAnl5eaZb9+7dW100EbXe6cIKZF2sgtLZCcO6+0tdDhER5JY03rZtm9n9FStWICAgAPv378fw4cOv+dyAgAB4eXlZXCARWZdxFs2Qrr5wU1n0K4CIqE3c0JiR0tL6S4/7+Phct22/fv0QHByM+Ph47Nq165pttVotysrKzG5EZB0/NYwX4SkaIrIVrQ4jQgjMmTMHt956K2JjY5ttFxwcjOXLlyMpKQkbNmxAVFQU4uPjsXv37mafk5iYCI1GY7qFhYW1tkwiusLFylocyKq/Hs1tnNJLRDZCJoQQrXnijBkzsHnzZvz6668IDQ216LkJCQmQyWTYtGlTk49rtVpotVrT/bKyMoSFhaG0tBSenp6tKZeIACTtP48Xvz2ImGBPbH1+mNTlEFEHV1ZWBo1Gc93P71b1jDz33HPYtGkTdu3aZXEQAYDBgwfj1KlTzT6uUqng6elpdiOiG/dTBk/REJHtsWj0mhACzz33HDZu3Ijk5GRERka2aqfp6ekIDg5u1XOJqHVqdQbsPlkMgKuuEpFtsSiMzJgxA19//TX+97//wcPDA/n5+QAAjUYDtVoNAJg/fz5ycnKwatUqAMDixYsRERGBXr16oba2FmvWrEFSUhKSkpKs/FaI6Fr2nb2ACq0O/h4q9O6kkbocIiITi8LI0qVLAQAjR440275ixQo8/vjjAIC8vDxkZWWZHqutrcXcuXORk5MDtVqNXr16YfPmzRg/fvyNVU5EFjEudHZbVACcuOoqEdmQVg9gbU8tHQBDRE0TQmDYol04f6kayx8dgDG9gqQuiYgcQJsOYCUi+3KyoALnL1VDKXfCrd39pC6HiMgMwwiRAzBeGO+Wrr5wVXLVVSKyLQwjRA7AuOoqZ9EQkS1iGCHq4IortEjPLgEAxHN9ESKyQQwjRB3croxCCAH0CvFEsEYtdTlERI0wjBB1cMYpvTxFQ0S2imGEqAPT6vT45VQRAC4BT0S2i2GEqAPbe+YiKmv1CPBQITaEq64SkW1iGCHqwC7PouGqq0RkuxhGiDooIcTl8SLRHC9CRLaLYYSog8rIL0dOSTVUcifc0o2rrhKR7WIYIeqgjKdobu3mB7XSWeJqiIiaxzBC1EHt5JReIrITDCNEHVBRuRYHz5cA4KqrRGT7GEaIOiDjqqu9QzUI9HSRuhwiomtiGCHqgIxX6eUsGiKyBwwjRB1MTZ0ev5wqBsBTNERkHxhGiDqYPWcuoLpOj2CNC3qFeEpdDhHRdTGMEHUwxim9t0UHQCbjqqtEZPsYRog6ECEEfm6Y0ns7p/QSkZ1gGCHqQI7llSG3tAZqhTOGdPWVuhwiohZhGCHqQIzXorm1ux9cFFx1lYjsA8MIUQdinNJ7O2fREJEdYRgh6iAKympw6HwpAGBUNMMIEdkPhhGiDuLnjPpTNH3CvBDgwVVXich+MIwQdRDGKb23s1eEiOwMwwhRB1BTp8evp42rrnJKLxHZF4YRog7gt9PFqKkzIETjgphgD6nLISKyCMMIUQews2FKb3xMIFddJSK7wzBCZOeEEPg5o+EqvZzSS0R2iGGEyM4dySlDQZkWrkpnDO7CVVeJyP4wjBDZOeNCZ8O46ioR2SmGESI795PpFA1n0RCRfWIYIbJj+aU1OJJTBpkMuI3rixCRnWIYIbJjxl6RvmFe8HNXSVwNEVHrMIwQ2THjVXpv5ykaIrJjDCNEdqq6Vo/fTKuu8hQNEdkvhhEiO/Xr6WJodQZ08lIjKpCrrhKR/WIYIbJTpgvjxQRw1VUismsMI0R2yGAQ+CmjYbxIT44XISL7xjBCZIcO55SiqFwLd5UcN0dy1VUism8MI0R2yHiKZngPPyjl/DEmIvvG32JEdsh0ld5onqIhIvvHMEJkZ3JLqnEsrwxOMmAUV10log6AYYTIzhgHrvYP94aPm1LiaoiIbhzDCJGdMY4X4YXxiKijYBghsiNVtTr8/tcFAPXrixARdQQMI0R25JdTxajVGRDu44puAe5Sl0NEZBUMI0R25PIpGq66SkQdB8MIkZ0wGAR+zigCwKv0ElHHYlEYSUxMxMCBA+Hh4YGAgABMnDgRJ06cuO7zUlJSMGDAALi4uKBLly5YtmxZqwsmclQHz5eguEILD5UcAyN8pC6HiMhqLAojKSkpmDFjBvbu3YsdO3ZAp9NhzJgxqKysbPY5Z8+exfjx4zFs2DCkp6djwYIFmDVrFpKSkm64eCJH8lPDQmfDo/y56ioRdShySxpv27bN7P6KFSsQEBCA/fv3Y/jw4U0+Z9myZQgPD8fixYsBADExMUhLS8N7772HSZMmta5qIgeTU1KNpAPnAXAWDRF1PDf051VpaSkAwMen+S7jPXv2YMyYMWbbxo4di7S0NNTV1TX5HK1Wi7KyMrMbkaM6d6ES9y/bg7zSGoT7uGJ0zyCpSyIisqpWhxEhBObMmYNbb70VsbGxzbbLz89HYKD5YLvAwEDodDoUFxc3+ZzExERoNBrTLSwsrLVlEtm1v4oqcP8ne5BTUo0ufm5Y//RguKss6tAkIrJ5rQ4jM2fOxKFDh7B27drrtr16CqIQosntRvPnz0dpaanplp2d3doyiezWifxyPPDJXhSUadEj0B3rnh6MYI1a6rKIiKyuVX9iPffcc9i0aRN2796N0NDQa7YNCgpCfn6+2bbCwkLI5XL4+vo2+RyVSgWVStWa0og6hCM5pXj08324VFWHnsGeWDPtZl6Hhog6LIt6RoQQmDlzJjZs2ICff/4ZkZGR133OkCFDsGPHDrNt27dvR1xcHBQKhWXVEjmAA1mX8OCne3Gpqg59wryw9snBDCJE1KFZFEZmzJiBNWvW4Ouvv4aHhwfy8/ORn5+P6upqU5v58+fjscceM92fPn06zp07hzlz5uD48eP44osv8Pnnn2Pu3LnWexdEHcS+Mxfw6Gf7UF6jw8AIb6x5YhA0rgztRNSxWRRGli5ditLSUowcORLBwcGm2/r1601t8vLykJWVZbofGRmJLVu2IDk5GX379sXrr7+ODz74gNN6ia7y66liTF7xBypr9bilmy++nDoIHi4MIkTU8cmEcTSpDSsrK4NGo0FpaSk8PT2lLofI6n7OKMD0NQdQqzNgVJQ/lj4yAC4KZ6nLIiK6IS39/OYcQSKJbTuSj+fWHkCdXmBMz0B8+FA/qOQMIkTkOBhGiCT0vz9zMOebg9AbBBL6hODf9/eBwplLvRORY2EYIZLIN2nZmJd0CEIAk/qHYtG9veHs1PTaO0REHRnDCJEEVu89h1e+OwIAeOjmcLxxdyycGESIyEExjBC1s89+OYM3Nh8HAEy5JQL//4Seza5GTETkCBhGiNrRR7tO490fTwAAnhnZFf8YG8UgQkQOj2GEqB0IIfCfHSfxwc+nAQAv3N4Ds+K7MYgQEYFhhKjNCSGQuDUDy3efAQC8NC4a00d0lbgqIiLbwTBC1IYMBoFXvz+KVXvOAQAWJvTElFuuf00nIiJHwjBC1Eb0BoEFGw5jfVo2ZDLgzYk34aGbw6Uui4jI5jCMELUBnd6Aud8exHd/5sJJBrx3Xx/c0z9U6rKIiGwSwwiRldXpDXh+XTq2HM6H3EmGxX/viwm9Q6Qui4jIZjGMEFmREAIvbzyCLYfzoXR2wpKH+mFMryCpyyIismkMI0RWtHz3GaxPy4aTDPjo4f4Y3TNQ6pKIiGwer8hFZCXbjuTj7W0ZAICX7+zJIEJE1EIMI0RWcPh8KWavT4cQwKODO2PKLRFSl0REZDcYRohuUF5pNZ74MhU1dQaM6OGPhQm81gwRkSUYRohuQKVWh6kr01BYrkWPQHd8+FA/yJ35Y0VEZAn+1iRqJb1BYNbadBzPK4OfuxKfTx4ITxeF1GUREdkdhw4jZ4srMeOrA6jQ6qQuhezQm5uP46eMQqjkTlj+WBzCfFylLomIyC457NReg0Hg6dVpOFlQgaJyLVZOHQhXpcN+OchCq/eewxe/nQUA/Ov+Pugf7i1xRURE9sthe0acnGR4774+8HCR44/Mi3hiZRqqa/VSl0V2IOVkEV7ddBQAMHdMD66uSkR0gxw2jABA71AvrJo6CO4qOfacuYCnVqehpo6BhJp3sqAcM786AL1B4J7+nTBjVDepSyIisnsOHUYAoF+4N1ZOGQhXpTN+OVWM6Wv2Q6tjIKHGisq1mLIiFeVaHQZF+iDxnps4hZeIyAocPowAQFyED754fCBcFE5IPlGEGV8dQK3OIHVZZENq6vR4anUackqqEeHrik8eGQCV3FnqsoiIOgSGkQaDu/jii8kDoZI7YefxQjy39gDq9AwkVD/Yee63B5GeVQKNWoEvHh8Ibzel1GUREXUYDCNXGNrND58+FgelsxN+PFqA2ev+hI6BxOH9Z+dJ/HAoD3InGZY9MgBd/N2lLomIqENhGLnK8B7++OTRAVA4y7D5cB5e/PYg9AYhdVkkkaT95/Hhz6cBAG/dcxOGdPWVuCIioo6HC2s0YVR0AD5+eACeWbMf//szF3InJ7x7b284Odn3YMVanQG5JdXIuliF7EtVyLpYhdySGvi7qzAwwhsDIrwR4OEidZmN6PQGZOSX41huGboGuKF/uHe7DBzdd+YCXtpwCADwzMiuuD8urM33SUTkiGRCCJv/s7+srAwajQalpaXw9PRst/1uPZyHmWvToTcIPBAXhsR7brLpQCKEQFG51hQ0si82BI+LVTh/qRp5pdW4XidPZ19XxHX2wcAIb8RF+KCrv1u7zxi5UKHFgawSHMi6hAPnLuHQ+VJUXzHluou/G+6PC8M9/TohwLNtwlNmcSUmfvwbSqrqMC42CB891N+mjz0RkS1q6ec3w8h1fH8wF8+vS4dBAA/fHI43JsZKOp2zvKYO2RerkX2pPmRkX6xC9qX60HH+UhVq6q49xsVF4YQwb1eE+7gizMcVIV4uOH+pGqmZl5CRX4arvxt83JQY0NnbFE5iQzRQyq13ds/Y65GedckUQM5dqGrUzsNFjpggTxzJLUVVw+J0zk4yjIryx31xYbgtOgAKK12grrSqDn/7+DecKa5En1AN1j01BGolZ84QEVmKYcSKNqafx5xvDkII4PGhEe1+ifhzFyqxLOUv/Hi0ABcra6/Z1kkGBGvUCPNR1wcOb1eE+7oitCGA+Lkrm629rKYOB85dQlrmJaRmXsSf2SXQXjXFWSV3Qp8wL1M46R/uDY265ReHu16vh1H3AHf0D/dG/85e6B/uja7+7nBykqFCq8OWQ3lYn5aN/ecumdr7uStxT/9Q3DcgFN0DPVpcz9VqdQZM/uIP7DlzAZ281Ng4Y6hNnroiIrIHDCNW9m1aNv6RdAhCANNujcQ/74xp80BysqAcH+06je8P5pqdXvFxUyLMW41QH9fLgcPHFWE+aoR4qa3WQ1CrM+BobqkpnKSdu9QoDMlkQFSgBwZG+CAuwhsDI3wQ4qUGYFmvR98wr4bw4Y2+YV4tCjinCyvw7f5sJO3PQXGF1rS9X7gX7o8Lw4TewfCw4Cq6QgjMSzqEb9LOw03pjP8+MxQxwdJ8vxERdQQMI21g7R9ZmL/hMABg+oiumHdHVJsEksPnS7Fk1yn8eLTAtG1klD+eGt4FvUO94K6SZtyxEAJniiuRlnkRqZmXkJZ5EZlNhIsQjQtCvNQ4lldmOqVypW4B7ugffjl8dGvo9WitOr0BKSeK8E1aNn7OKISuIbm5KJww/qZgPBAXhkGRPtc9VstS/sLbWzPgJAM+nzwQo6IDWl0TERExjLSZ1Xsy8cr/6i+SNuu2bpgzJspqr/3H2YtYsus0dp8sAlDf63BHryDMGNUNsZ00VtuPNRWW12B/5iWkZl7C/nMXcSS3zGwqtIdKjr7hXugX7o3+4V7oF+YNjWvLeyssVVSuxcb08/gm7TxOF1aYtkf4uuK+uDBM6h+KIE3j0y7bjuRh+poDAIDX7uqFyUMj2qxGIiJHwTDShlb8dhavfX8MADBndA/Miu/e6tcSQmD3qWJ89PNp/JF5EUD9wMy7+4Tg2VFd0S2g9eMfpFCp1eFgdgnyy2oQ20lzw70erSWEQHp2Cb5Ny8b3B/NQodUBqB9TM7yHP+6PC0N8TABUcmccOl+C+z/Zg5o6AyYP6YzX7o5t93qJiDoihpE29unuM3hzy3EAwD/uiMKzIy27eqvBILD9WAE+2nUah3NKAQBKZyfcGxeK6cO7ItzX1eo1O6qqWh22Hs7HN2nZ2Hf2omm7t6sCd/fthM2H81BUrsXIKH989lgc5FYac0NE5OgYRtrBR7tO490fTwAA/jk+Bk8O73Ld5+j0BvxwKA8fJ5/GyYL60wguCic8NKgznhrepclTCGQ9mcWV+HZ/Nv67/zwKyi4Peo0O8sC304dYNOCViIiujWGknby/8xT+s/MkAGBhQk9MuSWyyXZanR4bDuRgWcpfphklHio5HhvaGVNviYSvu6rdaiZAbxDYfaoI36Zlo7iiFv95oC86NcwCIiIi62jp5zeXg79Bz9/eHTqDAR/+fBqvfX8McmcnPDq4s+nx6lo91qVmYfnuM8grrQFQf3rgiVsj8eiQCIvW6CDrqV8wLQCjojhjhohIagwjVjBndA/U6g34JOUMXvnuCOROMkzoHYzVe8/h81/O4kLD2hwBHio8NbwLHro5HK5KfumJiIgAnqaxGiEE3th8HJ//ehYyGeCulKO8YQZHqLca00d0xb0DQuGi4LLiRETkGHiapp3JZDK8fGcM9AaBlb9nolyrQ1d/Nzw7shvu6htitVVRiYiIOhqGESuSyWRYmNATsZ008HCR4/aYQDjzSq9ERETXxDBiZTKZDPcOCJW6DCIiIrvBcwdEREQkKYYRIiIikhTDCBEREUmKYYSIiIgkZXEY2b17NxISEhASEgKZTIbvvvvumu2Tk5Mhk8ka3TIyMlpbMxEREXUgFs+mqaysRJ8+fTBlyhRMmjSpxc87ceKE2YIn/v7+lu6aiIiIOiCLw8i4ceMwbtw4i3cUEBAALy8vi59HREREHVu7jRnp168fgoODER8fj127dl2zrVarRVlZmdmNiIiIOqY2DyPBwcFYvnw5kpKSsGHDBkRFRSE+Ph67d+9u9jmJiYnQaDSmW1hYWFuXSURERBK5oQvlyWQybNy4ERMnTrToeQkJCZDJZNi0aVOTj2u1Wmi1WtP9srIyhIWF2fSF8oiIiMhcSy+UJ8nU3sGDB+PUqVPNPq5SqeDp6Wl2IyIioo5JkjCSnp6O4OBgKXZNRERENsbi2TQVFRU4ffq06f7Zs2fx559/wsfHB+Hh4Zg/fz5ycnKwatUqAMDixYsRERGBXr16oba2FmvWrEFSUhKSkpKs9y6IiIjIblkcRtLS0jBq1CjT/Tlz5gAAJk+ejJUrVyIvLw9ZWVmmx2trazF37lzk5ORArVajV69e2Lx5M8aPH9/ifRqHtXBWDRERkf0wfm5fb3jqDQ1gbS/nz5/njBoiIiI7lZ2djdDQ0GYft4swYjAYkJubCw8PD8hkMqnLuS7j7J/s7GwOvrVxPFb2hcfLfvBY2Y+2PFZCCJSXlyMkJAROTs0PU7X4NI0UnJycrpmobBVnAtkPHiv7wuNlP3is7EdbHSuNRnPdNrxqLxEREUmKYYSIiIgkxTDSBlQqFRYuXAiVSiV1KXQdPFb2hcfLfvBY2Q9bOFZ2MYCViIiIOi72jBAREZGkGEaIiIhIUgwjREREJCmGESIiIpIUwwgRERFJimHEijIzM/HEE08gMjISarUaXbt2xcKFC1FbW2vWLisrCwkJCXBzc4Ofnx9mzZrVqA21vTfffBNDhw6Fq6srvLy8mmzDY2U7Pv74Y0RGRsLFxQUDBgzAL7/8InVJBGD37t1ISEhASEgIZDIZvvvuO7PHhRB49dVXERISArVajZEjR+Lo0aPSFOvAEhMTMXDgQHh4eCAgIAATJ07EiRMnzNpIeawYRqwoIyMDBoMBn3zyCY4ePYr//Oc/WLZsGRYsWGBqo9frceedd6KyshK//vor1q1bh6SkJLz44osSVu6Yamtrcd999+GZZ55p8nEeK9uxfv16zJ49G//85z+Rnp6OYcOGYdy4cWZXCCdpVFZWok+fPliyZEmTjy9atAj//ve/sWTJEqSmpiIoKAijR49GeXl5O1fq2FJSUjBjxgzs3bsXO3bsgE6nw5gxY1BZWWlqI+mxEtSmFi1aJCIjI033t2zZIpycnEROTo5p29q1a4VKpRKlpaVSlOjwVqxYITQaTaPtPFa2Y9CgQWL69Olm26Kjo8VLL70kUUXUFABi48aNpvsGg0EEBQWJt99+27StpqZGaDQasWzZMgkqJKPCwkIBQKSkpAghpD9W7BlpY6WlpfDx8THd37NnD2JjYxESEmLaNnbsWGi1Wuzfv1+KEqkZPFa2oba2Fvv378eYMWPMto8ZMwa///67RFVRS5w9exb5+flmx06lUmHEiBE8dhIrLS0FANPnk9THimGkDf3111/48MMPMX36dNO2/Px8BAYGmrXz9vaGUqlEfn5+e5dI18BjZRuKi4uh1+sbHYvAwEAeBxtnPD48drZFCIE5c+bg1ltvRWxsLADpjxXDSAu8+uqrkMlk17ylpaWZPSc3Nxd33HEH7rvvPkybNs3sMZlM1mgfQogmt5NlWnOsroXHynZc/TXncbAfPHa2ZebMmTh06BDWrl3b6DGpjpW8zffQAcycORN///vfr9kmIiLC9P/c3FyMGjUKQ4YMwfLly83aBQUFYd++fWbbLl26hLq6ukaJlCxn6bG6Fh4r2+Dn5wdnZ+dGf50VFhbyONi4oKAgAPV/dQcHB5u289hJ57nnnsOmTZuwe/duhIaGmrZLfawYRlrAz88Pfn5+LWqbk5ODUaNGYcCAAVixYgWcnMw7n4YMGYI333wTeXl5pgO+fft2qFQqDBgwwOq1OxpLjtX18FjZBqVSiQEDBmDHjh3429/+Ztq+Y8cO3H333RJWRtcTGRmJoKAg7NixA/369QNQPwYoJSUF77zzjsTVORYhBJ577jls3LgRycnJiIyMNHtc8mPV5kNkHUhOTo7o1q2buO2228T58+dFXl6e6Wak0+lEbGysiI+PFwcOHBA7d+4UoaGhYubMmRJW7pjOnTsn0tPTxWuvvSbc3d1Fenq6SE9PF+Xl5UIIHitbsm7dOqFQKMTnn38ujh07JmbPni3c3NxEZmam1KU5vPLyctPPDgDx73//W6Snp4tz584JIYR4++23hUajERs2bBCHDx8WDz74oAgODhZlZWUSV+5YnnnmGaHRaERycrLZZ1NVVZWpjZTHimHEilasWCEANHm70rlz58Sdd94p1Gq18PHxETNnzhQ1NTUSVe24Jk+e3OSx2rVrl6kNj5Xt+Oijj0Tnzp2FUqkU/fv3N01JJGnt2rWryZ+jyZMnCyHqp4wuXLhQBAUFCZVKJYYPHy4OHz4sbdEOqLnPphUrVpjaSHmsZA1FEhEREUmCs2mIiIhIUgwjREREJCmGESIiIpIUwwgRERFJimGEiIiIJMUwQkRERJJiGCEiIiJJMYwQERGRpBhGiIiISFIMI0RERCQphhEiIiKS1P8DUBMmbDF7H2cAAAAASUVORK5CYII=",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(satisf_delivery_trend['low_bound'], satisf_delivery_trend['avg_stars'])\n",
    "plt.title('Satisfaction Trend vs. Delivery Time')\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "df.to_csv(r'C:\\Users\\matte\\OneDrive\\Desktop\\GitHub\\data\\ecom_brazil\\tableau_good\\review_score_vs_delivery.csv',\n",
    "          index=False)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Brazil and the tendency to buy locally"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 78,
   "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>case</th>\n",
       "      <th>frequency</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>km 0</td>\n",
       "      <td>5863</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>local</td>\n",
       "      <td>34908</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>not local</td>\n",
       "      <td>71879</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "        case  frequency\n",
       "0       km 0       5863\n",
       "1      local      34908\n",
       "2  not local      71879"
      ]
     },
     "execution_count": 78,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = '''\n",
    "SELECT\n",
    "\tCASE\n",
    "\t\t WHEN sellers.seller_city = customers.customer_city THEN 'km 0'\n",
    "\t\t WHEN sellers.seller_state = customers.customer_state THEN 'local'\n",
    "\t\t ELSE 'not local'\n",
    "\tEND,\n",
    "\tCOUNT(*) AS frequency\n",
    "FROM orders\n",
    "LEFT JOIN order_items AS items ON items.order_id = orders.order_id\n",
    "JOIN customers ON customers.customer_id = orders.customer_id\n",
    "JOIN sellers ON sellers.seller_id = items.seller_id\n",
    "GROUP BY CASE\n",
    "\t\t \tWHEN sellers.seller_city = customers.customer_city THEN 'km 0'\n",
    "\t\t\tWHEN sellers.seller_state = customers.customer_state THEN 'local'\n",
    "\t\t \tELSE 'not local'\n",
    "\t\t END;\n",
    "'''\n",
    "\n",
    "pd.read_sql(query, engine)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 79,
   "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>avg_distance</th>\n",
       "      <th>median_distance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>597.187869</td>\n",
       "      <td>431.753988</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   avg_distance  median_distance\n",
       "0    597.187869       431.753988"
      ]
     },
     "execution_count": 79,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = '''\n",
    "WITH distances AS(\n",
    "    SELECT\n",
    "        ACOS(\n",
    "            GREATEST(\n",
    "                LEAST(\n",
    "                    SIN(RADIANS(geo_sellers.geolocation_lat)) * SIN(RADIANS(geo_cust.geolocation_lat)) +\n",
    "                    COS(RADIANS(geo_sellers.geolocation_lat)) * COS(RADIANS(geo_cust.geolocation_lat)) *\n",
    "                    COS(RADIANS(geo_cust.geolocation_lng) - RADIANS(geo_sellers.geolocation_lng)),\n",
    "                    1\n",
    "                ),\n",
    "                -1\n",
    "            )\n",
    "        ) * 6371 AS distance\n",
    "    FROM orders\n",
    "    LEFT JOIN order_items AS items ON items.order_id = orders.order_id\n",
    "    JOIN customers ON customers.customer_id = orders.customer_id\n",
    "    JOIN sellers ON sellers.seller_id = items.seller_id\n",
    "    JOIN geolocation as geo_sellers ON geo_sellers.geolocation_zip_code_prefix = sellers.seller_zip_code_prefix\n",
    "    JOIN geolocation as geo_cust ON geo_cust.geolocation_zip_code_prefix = customers.customer_zip_code_prefix\n",
    ")\n",
    "\n",
    "SELECT\n",
    "    AVG(distance) AS avg_distance,\n",
    "    PERCENTILE_CONT(0.5) WITHIN GROUP(ORDER BY distance) AS median_distance\n",
    "FROM distances;\n",
    "'''\n",
    "\n",
    "pd.read_sql(query, engine)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Most sold categories"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 81,
   "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>category</th>\n",
       "      <th>frequency</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>bed_bath_table</td>\n",
       "      <td>11115</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>health_beauty</td>\n",
       "      <td>9670</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>sports_leisure</td>\n",
       "      <td>8641</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>furniture_decor</td>\n",
       "      <td>8334</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>computers_accessories</td>\n",
       "      <td>7827</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                category  frequency\n",
       "0         bed_bath_table      11115\n",
       "1          health_beauty       9670\n",
       "2         sports_leisure       8641\n",
       "3        furniture_decor       8334\n",
       "4  computers_accessories       7827"
      ]
     },
     "execution_count": 81,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = '''\n",
    "SELECT\n",
    "\ttranslat.product_category_name_english AS category,\n",
    "\tCOUNT(*) AS frequency\n",
    "FROM products\n",
    "JOIN product_category_name_translation AS translat ON translat.product_category_name = products.product_category_name\n",
    "JOIN order_items AS items ON items.product_id = products.product_id\n",
    "GROUP BY translat.product_category_name_english\n",
    "ORDER BY COUNT(*) DESC;\n",
    "'''\n",
    "\n",
    "pd.read_sql(query, engine).head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Orders with just one category"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 82,
   "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>one_cat_orders</th>\n",
       "      <th>tot_orders</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>8914</td>\n",
       "      <td>9673</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   one_cat_orders  tot_orders\n",
       "0            8914        9673"
      ]
     },
     "execution_count": 82,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "query = '''\n",
    "WITH sold_togheter AS(\n",
    "\tSELECT\n",
    "\t\ttranslat.product_category_name_english AS category,\n",
    "\t\titems.product_id AS product_id,\n",
    "\t\titems.order_id AS order_id,\n",
    "\t\tMAX(items.order_item_id) OVER (PARTITION BY items.order_id) AS num_items\n",
    "\tFROM products\n",
    "\tJOIN product_category_name_translation AS translat ON translat.product_category_name = products.product_category_name\n",
    "\tJOIN order_items AS items ON items.product_id = products.product_id\n",
    "\tORDER BY items.order_item_id DESC\n",
    "),\n",
    "\n",
    "categories_sold_togheter AS(\n",
    "\tSELECT\n",
    "\t\torder_id,\n",
    "\t\tcategory,\n",
    "\t\tnum_items,\n",
    "\t\tCOUNT(*) AS num_category\n",
    "\tFROM sold_togheter\n",
    "\tWHERE num_items > 1\n",
    "\tGROUP BY order_id, category, num_items\n",
    "\tORDER BY num_items DESC\n",
    ")\n",
    "\n",
    "SELECT\n",
    "\tSUM(CASE WHEN num_items=num_category THEN 1 ELSE 0 END) AS one_cat_orders,\n",
    "\tCOUNT(DISTINCT order_id) as tot_orders\n",
    "FROM categories_sold_togheter;\n",
    "'''\n",
    "\n",
    "pd.read_sql(query, engine)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "**Findings:**\n",
    "The vast majority of orders (with more than 1 product) in Brazil are within the same product category."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "base",
   "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": "3.11.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}