Conteudo 6 - Processo de modelagem com Maquina de aprendizado.ipynb

{
    "nbformat_minor": 2, 
    "cells": [
        {
            "source": "# Processo de modelagem com Machine Learning ", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "<img src = 'https://frnsys.com/assets/hosny/disparity.gif' width =600>", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "\nNeste notebook vamos apresentar conceitos e c\u00f3digos referentes a um processo de modelagem do inicio ao fim, testando o modelo desenvolvido. \n\nAinda que simples serve como uma refer\u00eancia para modelagens futuras.\n\nO processo mais adequado e completo ser\u00e1 apresentado durante as aulas e trata-se do CRISP-DM\n", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "## Processo b\u00e1sico de modelagem\n\n\u00c9 composto por 5 passos:\n\n* Fazer a pergunta certa\n* Preparar os dados\n* Selecionar o algoritmo\n* Treinar o modelo\n* Testar o modelo", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "### Preparando os dados", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "\nAntes de iniciar a prepara\u00e7\u00e3o dos dados necessitamos importar um conjunto de bibliotecas referente a cria\u00e7\u00e3o de estruturas de dados, entendimento num\u00e9rico dos tipos de dados al\u00e9m de bibliotecas gr\u00e1ficas.\n", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 11, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "import numpy as np\nimport pandas as pd\nimport seaborn as sns\nimport matplotlib.pyplot as plt\n%matplotlib inline\n\nplt.style.use('seaborn')"
        }, 
        {
            "source": "Agora vamos carregar uma base de dados sobre consumo de cerveja e analisar o quanto o clima impacta nisto", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "<img src = 'https://media.giphy.com/media/42dsvcMDP3diU/giphy.gif' width=500>", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "df_beer = pd.read_excel(\"data/beer_consumption.xlsx\")\ndf_beer.head(10)", 
            "cell_type": "raw", 
            "metadata": {}
        }, 
        {
            "execution_count": 4, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 4, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>42005.0</td>\n      <td>27.30</td>\n      <td>23.9</td>\n      <td>32.5</td>\n      <td>0.0</td>\n      <td>False</td>\n      <td>25461.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>42006.0</td>\n      <td>27.02</td>\n      <td>24.5</td>\n      <td>33.5</td>\n      <td>0.0</td>\n      <td>False</td>\n      <td>28972.0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>42007.0</td>\n      <td>24.82</td>\n      <td>22.4</td>\n      <td>29.9</td>\n      <td>0.0</td>\n      <td>True</td>\n      <td>30814.0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>42008.0</td>\n      <td>23.98</td>\n      <td>21.5</td>\n      <td>28.6</td>\n      <td>1.2</td>\n      <td>True</td>\n      <td>29799.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>42009.0</td>\n      <td>23.82</td>\n      <td>21.0</td>\n      <td>28.3</td>\n      <td>0.0</td>\n      <td>False</td>\n      <td>28900.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "      data  temp_avg  temp_min  temp_max  precip weekend  beer_consumption\n0  42005.0     27.30      23.9      32.5     0.0   False           25461.0\n1  42006.0     27.02      24.5      33.5     0.0   False           28972.0\n2  42007.0     24.82      22.4      29.9     0.0    True           30814.0\n3  42008.0     23.98      21.5      28.6     1.2    True           29799.0\n4  42009.0     23.82      21.0      28.3     0.0   False           28900.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "# The code was removed by Watson Studio for sharing."
        }, 
        {
            "execution_count": 5, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 5, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>count</th>\n      <td>365.000000</td>\n      <td>361.000000</td>\n      <td>363.000000</td>\n      <td>362.000000</td>\n      <td>365.000000</td>\n      <td>365.000000</td>\n    </tr>\n    <tr>\n      <th>mean</th>\n      <td>42187.000000</td>\n      <td>21.101385</td>\n      <td>17.396970</td>\n      <td>26.480663</td>\n      <td>5.196712</td>\n      <td>25401.367123</td>\n    </tr>\n    <tr>\n      <th>std</th>\n      <td>105.510663</td>\n      <td>3.546618</td>\n      <td>2.969405</td>\n      <td>4.758531</td>\n      <td>12.417844</td>\n      <td>4399.142703</td>\n    </tr>\n    <tr>\n      <th>min</th>\n      <td>42005.000000</td>\n      <td>0.000000</td>\n      <td>0.000000</td>\n      <td>0.000000</td>\n      <td>0.000000</td>\n      <td>14343.000000</td>\n    </tr>\n    <tr>\n      <th>25%</th>\n      <td>42096.000000</td>\n      <td>18.920000</td>\n      <td>15.200000</td>\n      <td>23.725000</td>\n      <td>0.000000</td>\n      <td>22008.000000</td>\n    </tr>\n    <tr>\n      <th>50%</th>\n      <td>42187.000000</td>\n      <td>21.360000</td>\n      <td>17.900000</td>\n      <td>26.900000</td>\n      <td>0.000000</td>\n      <td>24867.000000</td>\n    </tr>\n    <tr>\n      <th>75%</th>\n      <td>42278.000000</td>\n      <td>23.280000</td>\n      <td>19.550000</td>\n      <td>29.400000</td>\n      <td>3.200000</td>\n      <td>28631.000000</td>\n    </tr>\n    <tr>\n      <th>max</th>\n      <td>42369.000000</td>\n      <td>28.860000</td>\n      <td>24.500000</td>\n      <td>36.500000</td>\n      <td>94.800000</td>\n      <td>37937.000000</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "               data    temp_avg    temp_min    temp_max      precip  \\\ncount    365.000000  361.000000  363.000000  362.000000  365.000000   \nmean   42187.000000   21.101385   17.396970   26.480663    5.196712   \nstd      105.510663    3.546618    2.969405    4.758531   12.417844   \nmin    42005.000000    0.000000    0.000000    0.000000    0.000000   \n25%    42096.000000   18.920000   15.200000   23.725000    0.000000   \n50%    42187.000000   21.360000   17.900000   26.900000    0.000000   \n75%    42278.000000   23.280000   19.550000   29.400000    3.200000   \nmax    42369.000000   28.860000   24.500000   36.500000   94.800000   \n\n       beer_consumption  \ncount        365.000000  \nmean       25401.367123  \nstd         4399.142703  \nmin        14343.000000  \n25%        22008.000000  \n50%        24867.000000  \n75%        28631.000000  \nmax        37937.000000  "
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.describe()"
        }, 
        {
            "execution_count": 6, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 6, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "data                365\ntemp_avg            361\ntemp_min            363\ntemp_max            362\nprecip              365\nweekend             363\nbeer_consumption    365\ndtype: int64"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.count()"
        }, 
        {
            "execution_count": 7, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 7, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "<matplotlib.axes._subplots.AxesSubplot at 0x7f5cd1889668>"
                    }, 
                    "output_type": "execute_result"
                }, 
                {
                    "output_type": "display_data", 
                    "data": {
                        "image/png": "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\n", 
                        "text/plain": "<matplotlib.figure.Figure at 0x7f5cd2b13ba8>"
                    }, 
                    "metadata": {}
                }
            ], 
            "source": "sns.heatmap(df_beer.corr())"
        }, 
        {
            "execution_count": 8, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 8, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>data</th>\n      <td>1.000000</td>\n      <td>-0.144410</td>\n      <td>-0.127932</td>\n      <td>-0.029035</td>\n      <td>0.007490</td>\n      <td>0.043541</td>\n    </tr>\n    <tr>\n      <th>temp_avg</th>\n      <td>-0.144410</td>\n      <td>1.000000</td>\n      <td>0.678633</td>\n      <td>0.735339</td>\n      <td>0.026834</td>\n      <td>0.503227</td>\n    </tr>\n    <tr>\n      <th>temp_min</th>\n      <td>-0.127932</td>\n      <td>0.678633</td>\n      <td>1.000000</td>\n      <td>0.531704</td>\n      <td>0.094701</td>\n      <td>0.345033</td>\n    </tr>\n    <tr>\n      <th>temp_max</th>\n      <td>-0.029035</td>\n      <td>0.735339</td>\n      <td>0.531704</td>\n      <td>1.000000</td>\n      <td>-0.071084</td>\n      <td>0.587027</td>\n    </tr>\n    <tr>\n      <th>precip</th>\n      <td>0.007490</td>\n      <td>0.026834</td>\n      <td>0.094701</td>\n      <td>-0.071084</td>\n      <td>1.000000</td>\n      <td>-0.193784</td>\n    </tr>\n    <tr>\n      <th>beer_consumption</th>\n      <td>0.043541</td>\n      <td>0.503227</td>\n      <td>0.345033</td>\n      <td>0.587027</td>\n      <td>-0.193784</td>\n      <td>1.000000</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "                      data  temp_avg  temp_min  temp_max    precip  \\\ndata              1.000000 -0.144410 -0.127932 -0.029035  0.007490   \ntemp_avg         -0.144410  1.000000  0.678633  0.735339  0.026834   \ntemp_min         -0.127932  0.678633  1.000000  0.531704  0.094701   \ntemp_max         -0.029035  0.735339  0.531704  1.000000 -0.071084   \nprecip            0.007490  0.026834  0.094701 -0.071084  1.000000   \nbeer_consumption  0.043541  0.503227  0.345033  0.587027 -0.193784   \n\n                  beer_consumption  \ndata                      0.043541  \ntemp_avg                  0.503227  \ntemp_min                  0.345033  \ntemp_max                  0.587027  \nprecip                   -0.193784  \nbeer_consumption          1.000000  "
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.corr()"
        }, 
        {
            "source": "N\u00e3o foram encontrados colunas correlacionadas o suficiente para remov\u00ea-las.", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "### Compreendendo rela\u00e7\u00f5es e correla\u00e7\u00f5es de vari\u00e1veis", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "sns.lmplot()", 
            "cell_type": "raw", 
            "metadata": {}
        }, 
        {
            "execution_count": 13, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 13, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "<seaborn.axisgrid.FacetGrid at 0x7f5cc1d6ca90>"
                    }, 
                    "output_type": "execute_result"
                }, 
                {
                    "output_type": "display_data", 
                    "data": {
                        "image/png": "iVBORw0KGgoAAAANSUhEUgAAAWAAAAFgCAYAAACFYaNMAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMS4wLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvpW3flQAAIABJREFUeJzsvXl4HOWVLv5+Vd3VrZZa+2pjC8sLyGBsMoANhIAcsMGGISaDY0/ClkmMITNAmN8QMD/uZCYTtpAJTjIEcnNhyGSuHWYGSLC8yIBMHBazeiGWY8C2bNmSrF2t3qq76rt/fP1VV3VXr+rWRr3Pw/PgVndt3fXW+c55z3sIpZTCggULFiyMOYTxPgALFixY+LzCImALFixYGCdYBGzBggUL4wSLgC1YsGBhnGARsAULFiyMEywCtmDBgoVxgi2fGw8Gg/j6178OWZahKAqWL1+Ou+66CydOnMC9996LoaEhzJ8/H48//jgkSUq4nZ4eTz4PEwBQVubCwIAv7/sZT1jnOHXweTjPqXSOVVVu09fzGgFLkoTnn38ev//97/Hyyy9j9+7d2Lt3L5544gnceuutaGlpQXFxMf77v/87n4eRFmw2cbwPIe+wznHq4PNwnp+Hc8wrARNCUFhYCAAIh8MIh8MghOCdd97B8uXLAQCrVq3Ca6+9ls/DsGDBgoUJibymIABAURTccMMNOH78OP76r/8aM2bMQHFxMWw2tuva2lp0d3cn3UZZmWtMnoaJlglTCdY5Th18Hs5zqp9j3glYFEX87ne/w/DwML7zne/gyJEjce8hhCTdxljkgaqq3GOSax5PWOc4dfB5OM+pdI7jkgPWo7i4GIsXL8bevXsxPDyMcDgMAOjq6kJ1dfVYHYYFCxYsTBjklYD7+/sxPDwMAAgEAnjrrbcwe/ZsLF68GDt27AAAvPTSS1i6dGk+D8OCBQsWJiTymoI4ffo07r//fiiKAkoprr76ajQ1NWHOnDn47ne/iyeffBKNjY248cYb83kYFixYsDAhQSaDHeVY5IGmUr4pEaxznDr4PJznVDrHcc8BW7BgwYIFIywCtmDBgoVxQt5laBYsWLAwWdDaKmLTJjva2wXU16tYuzaEpiYlb/uzCNiCBQsWwMj34Ycd2r+PHhUi/w7mjYStFIQFCxYsANi0yW76+ubN5q/nAhYBW7BgwQKA9nZzOkz0ei5gEbAFCxYsAKivVzN6PRewCNiCBQsWAKxdGzJ9fc0a89dzAasIZ8HCBEE2FXj+mQMHBAQCBE4nxYIF+a/eT0Ww6xXE5s3R72DNGksFYcFCUuiJa9484IYbxElHPtlU4PlnPB6gs5MvZgkoBY4ezW/1fiIgH5KxpiZlTK+ZlYKwMKnBSejoUQGqCnz6KfDwww60tqbvH93aKmLdOieWL3dh3TpnRp/NFbKpwPPP9Pcb7Vz5v/NZvR9vxH7v/IE1Ht/daGARsIVJjdFKhybKjZxNBZ7/LRg0ErAsk5Sf1WMiPIAyxXhIxvIBi4AtTGqMVjo0UW7kbCrw/G8Oh9FPS5Joys9yTJQHUKYYD8lYPjC5jtaChRiMVjo0UW7kbCrw/DPl5UYC5v9Op3o/UR5AmWI8JGP5gFWEszCpsXZtyFC84khXOlRfr+LoUQEeD8udBoMEDgfF/PljeyNnU4HXf0YQYFBBpFu9nygPoFi0top48UXg8GGXaYHN7Hv3eIDuboLly80/My6gFKDhhH+2CNjCpEYscc2dC6xalX71f+3aEB54wKFTEbCcalcXQWvr2KopsqnAj7Zqzx9AZq+PF3haxGaDIS2iV3XEfu9OJ4XHQ+D1svz3WPg4JIUaBgkPgYSHQcVCAOWmb7NSEBYmPZqaFDzzTADbt/uwaRMyuuGamhTU1lI4HBSqCoTD7Kbv7yfYuFHK41FPDIxH80EqpJsW0X/v1dUUbhPP8zFPpSheCIFOiP6jEEL9IEmiX8CKgC1McaSjFfX7CcrLKTo7CYRISBIMEuzdK455FDzWGI/mg1QYjSIkk8/kDFQBCQ+ziFfN7MFlEbCFKYt0mxvq61Xs2hVf9Zckis2b7VOagIH8Nx9k2jCRTVpkXFIpih9CeAgkPAIgu8luVgrCwpRFukvZtWtDcVpagKkJxrsYNdmRjcxtNIqQTD6TFagCEhqA4G+HGOgACXuQLfkCVgRsYQoj3WVpU5OCRYsUtLUJkGUCSaIoL2c5xXwXo8Z6AsNYI9lDMNF58rTISy+5cPgwMlaE5CWVovhBwsMQRkm4sbAI2MKURSbL0nvukSPpCePNlc9i1HhMYBhrZJubbWpSsHo10NPjS3tfOU+lUFWX25Vzt10dLAK2MGWRiUZ4PIpR2USHkw0TUeaWEkoAb7/hwfYtAZw6RTB9GsHKa0UsWZz778QiYAtTFpmS6lg7YU3UJohcYrSNMmMGLdodxp63w/jV0xIAVhc40SHg6aclAHLOSdgiYAtTGmNNqplgUkaHMUiVw56IMjcDlAAjXsUDQtl1b94S/8AAgOZmW+YETMMQvX8GMNv0zxYBW7AwTsg0Opxovsfp5rAn3EOQqiBhDyNeNRD355OnzFcgpxK8brZ92/BHkHpbIPW/DqXgTOCsFaZvtQjYgoU0kC/z73Sjw1iy477H41mwm3Q5bDUIEhoyRLtmmD5NxYmOeLKdNi3JyoRSiCMfw9G7A1Lfq1BtpZArl2F4wbMIF52LygQfswjYgoUUMIv0HnjAgdpaCr+faIQMQCPpggIKSplBTjLCTjc6nIhkNyly2JQywg0NmUa7Zlh5bTiS8415fWVMWzGlEL2HWKTbtxMQHJArlmF4/i+gumaltS+LgC2MGSar5jWW/PgIoP5+ivp6qhEyALjd7O+HDrFGg7o6NSfysvEku0Tf24TOYauyZoaTLNo1A8vzymhutuHUKQHTpqlYuTLMXqcUou9TSH07IfW2AFSFXLkMI2f/BIprDkDiG3qSwSJgC2OCyax5jSU5PvKHTZ6ghtfcbmoYEdTfT+B2s/eMJlodL7JL9r1NOIUDj3bDwyCKf1SbWrJYMRTcBP8xSCda4OhtARQf5MqrMDLvYSiFjRmTrh4WAVsYE0yEJXS2EXgs+fG2ZT55gr/G7kNqaGvWk/RootV8kV2qa5Lse3vmmQAmhMJBlTUJGaG527cQ6IDUyyJdIdQPufJKeGc/hLB7AUBys/KwCNjCmGC884WtrSIeeMChma4fPUqwf7+ARx5JHYHHkp/DwUhWP4lCPxaI/x0wkvRootXR+h6bIZ1VSarvbdwUDpSCKCMszTDKaFcPIdgFqe9VRrqBU5ArlsI36+8RLj4fILkf02QRsIUxwXjnCzdulOJM1zs7medvU1PyGziW/ObPV9HVRQz+s3oy5taWsa+PNlrVk11VlRs9PaMjvnRWJeP9vcUhD9EukXs10hX9xxAqvwK+mXciXHIBQPJLkRYBW8gb9MvbggIKjwdxptljlS88eNA8kmtrY6+n01Cg/3drqxi39AagvVZby97LVRB8aT6RCpHprEomRJ5Xi3aHQZT0vSGSgYQGIPW1QurdAdF7CKHyLyEw/TaESpcAwtiZuFsEbCEviF3e8lExhYU0jpTS3V6+iCvVUjzRvhPJyrLdz1gjneh2XDvZ1BCI3MfSDDmIdknYA3t/K6TenbB59iFUegkCdV9DqPQSQHTm4IAzh0XAFvICs+Wt2w3U1FA884z5kj8R0eWCuBobVezdG5/Da2xUU/oG54o0J0IhUo90o9sxzfNSysb6hIeBEQIhNDK67SleSP1vMNId/gDhkgshV1+HkbMeB8SC3BzzKGARsIW8INOiWzKSzQVxNTWFsX+/gECAReJOJ0VNDcXdd8t49FHz3v8DBwS8+64T/f1sUjL3CM503xzJrsl4pCYmlE+DbohldI5aUXbbUvywD/wRjt6dsA29g7B7EeTK5RiZ+wPAluU28wSLgC3kBZkWb5KRbDYKCj2hOZ0U3d0ENTUU/f1MGma3QyObTZvij1U/ph6IFu0AFW53duqNRNfE6aQ5T02kS+jj6tOg5XY9IIp3dNtSZdgH34bUuwP2gTehFJ2DYOUyeGc/CGovyc3x5gETqGfQwlRCpiNikpFsItJO9HrsGJy2NkFTQNTXU8ydq6K+nqKtTUx4rP39TGaml5fx15PtOxkSXZNEyHaibzZjgMYUqgwi90LwH4UQ7MqefNUQ7ANvovCTf0Tp+8vhPPUbhN2LMHT+S/Cc8xTkmq9MDPJNohnOWwTc2dmJ++67D729vRAEAatXr8Ytt9yCn/3sZ3jhhRdQXl4OALj33ntx+eWX5+swLIwT+PJ240ZJUxqcfXZ2QxXXrMmsEh8bTfMoVt+VBhi1rLFL8ZERAZQCoRAQCACCAIhitLEiGxVAoiV/ohRIthrpiZZrBpC7LjWqwDb0AaS+Fkh9rVCdMxCsXAbfzO+AOqpzd7yjACUCILhAxQJQ0QUI8b4SHHkjYFEUcf/99+Occ87ByMgIvvrVr+LSSy8FANx66634m7/5m3zt2sIEgtdLMHMmIz2fjyRcWicrCGWaq4wlLt4Yoe9KA+Kr/frtfeUrBfjoIxYx2u2AogDhMFBcTLFhQ2apgdTdZrnV2o5304sBmieDJ3slA1Vh8+yLmN68BlWqZk5j5/0GqrMut8ebFUiEbAtABVdGioq8EXB1dTWqq9kTqaioCA0NDeju7s7X7ixMQGQSiaUi2UxylbHRNG+M0HelAcm1rFT3VlFk/wHAvHlqxuSbKr+ba62t00nR1iYgGDQWD7Mh9KyKg7noUqMUGNgH19HfQerbCdVWzEj33F9BLZiZ3TZzCCo4QMVCULEAEAqy9oMglNLcjfhMgI6ODnzjG9/Ali1b8Nxzz+Gll15CYWEhzj33XNx///0oKUmepwmHFdhsEyR/ZSElWlqAZ58FtmwBJAmorASKi6N/F0Vgz57Rb//IEaChAfjmN4Fly4x/37DB+JnhYWDaNMDvZ5+57TbjZ2Jx0UXA4CDQ2wsEg4DDwc6jrCx67KmOAwDWrGHevbGYOxfYtMl4zM89F91W7PGlsy/+vr/9W+DkSePr06cDP/959DPpbM/sOgLAww8nuHZqGAgNsf/UsMkbUoBSYLgNOLWV/SdIwLQV7D/3nMy3l0sQEbC5AFshILoAITexa94J2Ov14qabbsL69euxbNky9Pb2oqysDIQQbNy4EadPn8YjjzySdBs9PZ58HiIA3tqZ//2MJ8biHPURX3t7VEVQV6dqEq6GBjVi5DK67evB0wL8HM061RJFbmZR3qZNduzfL2hKCB5JLlzIjj3VcXAsX+6CahJ4iiKwfXt6XV1m+7LZRNx3ny/unNatc+LoUUFTccgyASEUTicwcyZFfb2K+fMVNDfHr05ij51vKxZx35/ihRAajhTTMqcT0fcZpN4dkHp3AlSBXHkV5MplKDnjC+gfGKU6Imtkn1YwQ1WV2/T1vMrQQqEQ7rrrLlx33XVYFnlkVlZGveFvvPFGrF+/Pp+HYGGMoU876D0R9AWwVEvrZMvedNMa6aYsEqUIGhsVU+8IfuzpHgdPh+hlbQ4Hxfz5RlbO5pw3bpSwaRM1fIbned1uZo3p8VB0dgrw+wFVZd7Fu3aJBk1zomNPmktWQzpPhsyjXcF/nOV0e1tAlBHIFVdiZO4PoBSdE13Oj8LmMRuwtIIrklZwjcn+80bAlFI8+OCDaGhowG233aa9fvr0aS03/Oqrr2Lu3Ln5OgQL4wD9TctucFWLxBoalJRC/1Q501wXmBKR2+7dNtTVRY9dklgEzKRrobSPY+3aEB54wBFH5l1dBK2tYlqdfmb7Gh4Gjh8XUVBADe5utbUUXV1EI3tZZit7QtiKpLycvb+/HwZFiNmxx+bSCVS4pBHMnzsA0T9oev7JIARORQppLRDkXsgVV8I3ewPC7vNyZu+YCSixMZUCj3JzlFbIBHnb4wcffIDf/e53mDdvHq6//noATHK2ZcsWHDp0CAAwffp0/PM//3O+DsHCOCD2puWRWEODklbaIVVkmWt3rkREOjBAMHcu1aLI/n6Czk4BAwOMONM9jqYmJdIAQg1E7nZHzymbc+7oAGSZyeIEAVAURrjhsIqeHvZeRWH/AYDNFo3iBYHGKULMjp0XBx02P4ocQ3DZR0CIilV/KSe8nrEgwdOQ+l6Fo3cHhEAHs3c887sIF38hL/aOyZCJPGyskDcCvuCCC/DnP/857nVL8zu1MdqKfqrIMteKgUREWlbGyImPH+KglHlDrFwZMv2c2XEw8yGKWMLj55TpOXs8rJjIV8iqCi3P3NkpYMYMFrl7PER7T2ylJ1YRAgCNjQrWrXOivV3ArPoQvrGmD//ywDC2Navxo3mSgMh9kPpeg9TXAtH3GUJll8M/43aESi4cU6exXOdx8wGrFdlCTmEmJ2tsZFHeo486UkqZUkWWufYvSETot9wio7nZbhgvBET9fdvaRGzYkN5xpDqnTM85GGRRbyyp8miXrzoOHyZQFNZMon8vpQR33hlEW5to+I6am+1w2PwoLxhEaHAEzz1NsX69jB/8c+prS0KDkPpfh9S7E+LIQYTKvojAtJsj9o5jF2lSwanL42YvDxsrjIkMbbSwVBC5wXicoz6/qS9ELVqk4J575DjCis2H8s8UFFAIAoHTSbFgQWISz+YcEykmWltF3H67U0sdOJ3MSjMYZMfx9NOBrAp9HFx1kOrvTz5px/PPSxgYICgro1BVwOMR4PUab11CmN2nKCKSjoAhArbbWeQ7f76Kl1/W6XOpgu/+rYz+Lg9sIksveL3A0BCLoC+6UMHKa+MjX2bvuCti77gXodKLIVcuQ6j00pxEm+XlRejvT+6GRgWJEa5QwORh45BLTgfjooKwMHmRK3cunt/0eICODgGKwpbL77/PRgTFjgTSR3sHDjAZmNNJMTjIbywCSoGjR3Pno5vM2/eKKxRNxWCWikjnGNJpMjH7OwBcfrkLhw4JWis0KwoCpaUsB8yvpyAANhtFYSEjToCRbyjEiHfGjKgM8O67ZbS2ivif34YxcHoYc84cRlc7gcvF/u71QssjEwKc6BAiY9plLLnAA6n/D5D6WmAfeh+hkgsgV63AyFmPMgLMMygRAdEVId3xKZzlEpP76C3kBbk0Dud5zO5ugpAuPaoojNDMRgJxQuQ61PZ24zKSS9rGwtuApygSpSKSHUOqh1js3+fPV0ApcOwYuy5dXQQnTrDrp8/ziiJ7oJ1xhlGlUVdHIUmAy0W11+12oKCAorSUpTTWfi0AGx3G//63AGyiDKcN6OggGBwkoNRI4ABgt1NIYgB/MeMPqO3YhlL8EWH3QsiVy+Cd80+gNvPILncgWkqBFc7MfTMmKywCthCHXJq58Pwm9+HVJ7wUJToSyAycvPVThoGoIc5YeBvw6HT9eicIgUHFoD9GDk6qBw4I6Osj2nvNpmzoH3L79wvYscOmNawcPChoaQQ9FIV1F6oqsHChGiFvRTP1UdVoDpgX/UQR2N7cByE8BBIewUMPSbCJxuMuKaEYGiIoLGQKCbso45K5f8T1F2zF4ll/QNup+Xhlzwps/tNDuGJ5EZbMH92D7509Ipq32HDylIDp01RDioPnceGqgRIITfg87mhgEbCFOORSa8sjSErji0ahEAxRcSw4eeunDAPRCn6uBkMmm8TBX3e5gLIyNa55QX8MelLt6yNxHsJAYukZj7B5dB/70OFQFObOVlCAuKJfrKmPTQihUBpG45xBiIFoLvXkqfjvsbAQsNtkLDv/Hcy8YAcuP/s1tPfNwa7D1+BfXrof/d5KSBLFtGkUh54GADmlGiIR3tkjRlIaDB8fdGDnG+VwFrkw5ywn1qyNpIRsBQDJoqV5EsEiYAtxyKXWlkeQN91UADkiH9UHNM4ktRpO3vqOOiC6/M/FYMhE6ZZ9+0KGdl2Hg0ZywEYS1h+DnlT1BKrvAkwkPePv59E9f+jYbNCiYP0DrKwsPgfNrpcElzSCImkITjtrdf7L64y63enTVJzoYPsXiIJzat/HF2dvwxcbXkVBxTQc9i7HPf/5Evq8tTh1irvIsSiZo7nZljUBv/KKA/5QIfxyIXoGXDhxkl1/h4dCkCgeftgGIIjVq7Pa/KSCRcAW4pCJ1jadYl1Tk4IZM1ScOCEYikaiCJSWJhbh6ItTgsD0tFwFkavROYnSLc8/L6GyMnpsvKtPlglEkZrKzvSkqo/a9U0PiaRn/P08uucPncJC5uPQ3c22JYpAVRVFbS1LT2hpISWApZd6UPiAD1ubSVLd7sqVMt54+U+4bPY2XDKrBQO+Kuw+cg3eL/gPLDyvFrUAbhRENDeraG+3QZIoSkpYfpjjlEkUHQueZug4KaKqVsK110u49HIH9hws1/LZp3ujDyr9ddq82W4RsIXPJ9LV2mZSrFuwQAWliGvtXbAg8VQLPbH/8IdBAIwwjx0TNOLk++HvP3kSmD7dmbZqI1knnJ6AAUbCokgTmujoSVUfteubHvhDLPYhx99fXk416Z2isA42pnBgKweXi8LliuR2SRgD3cMQ/N0gKotyL17M/osDpRBHDkLq3YHltldx2V8WYdcn1+Chrb8GimZi5cowFuqIesliBUsWK3joIWjRsh7TpiVfDb29x4knNpYhEHIhEHKhvV/A+weBDSRouE5mqSVgnLyLxwEWAVswRTpmNpkU69auDUWkY2wuG/cjaGyM34dZgWrnTifCYaYHLi+nBrLft4+pBnih7+BBEfv3C3ESNzOk6oQze38i6ElV74PhclH09rLo3fjgiD7kFi5kD7ldu2zYu1eEJFGccQbbV2enAEmiUFWCYBDw9HsxrdSDstJhzJihauQbB0oh+j7RTG9ARMiVV2Gk8adQCudAFEXgUxtOnhTQvIVRQVy0fG3YkK/VXl9pzM1qvgqRAtqvNrsx4Iu/rps32w3XSb9S4Kml6HVO3Ko8HkNM8wGLgC1kjUyKdU1NCvbtC+GppyTIctTesbnZjoUL1ZhiUpTYuf6W549jC1sbN0rYt0+ELxKUEsImb3R0EFOJWyxSdcLFIlne2YxUeYcZX1rHrhJiSePgQRG1tarmPcGbKQqkANzOYRQ5PRCIioCXAMXxRAgAgu8IHBHTG6ghRrpnPQ6l8CwtAR9bCDNofWMiYUBGc7PNkNZYvISCCkWaPKz1jQIDIR44IKDIZABxe7tguE5er4DeXsS5s7HrbP6Az6VMcrxhEbCFrJFpse7gQdHUEyGZDSJXB6gqK0Lx+Wzd3QLcbhVtbQICJh4/qSRuHMnSLQsXqhm3POstJI8dE/Duu6I22FNvR5no4XDgQHSAqCiEUeAYRnHBMAoLZNTUqBgaIgiFCAiA9eujZCn4T7A5ab0tEMLDkCuuwsicf4JSdK6pjItHvHGvmxTXeDpCPwVC1bX5mhFiXx/TFSdSjegfPpl4N/Nra4ZxnXmXJSwCtpA1MjXG4drY2FE5yWwQg0GWC9UrAFSVmdF4YjqO9VI3MxP0REjWCZfpDR1LRiyXGy2gAeycPvhAxFe+UgC/nxiW0IEAUOTwoLhgGC6HF6FQ5OGjMqlYYSGF10uhKCJe+r89CP5pG5aevR1uezfkii/D13A/wu6FKVtyzaRogLG4RgUp0uIbiXJ3SabLfjNCLC+ncUNQAfPfRqbXeULNvBslLAIeY0yV3BWQmTFOa6uokS9gTCUsXGhugwiwHKHHE22t1ePECaYRNtMYUwrU1aVnc5LsO8n0+4olI/3xi2J0uKeqAu+9J2LGDPaweeIxCgftxVnTZPT0AGrERIdH/gBrERaCPVh61g5c+4VtOLPqKN45dhV+1Pz3uPS6hVjckH7Dgl6KxhFW7aiodUJ1lDDi1bX5muXld+1yoqKCGhpOONxuQBAoGhrUjFYQ6SDXlqTjCYuAxxBTKXfFkW70smmTPU7PC7AIMTYqis0RejwEUiRdyWVsAEtFVFdTHDsWTzw2W3KJG8C+jyeflLB3r6hF5PrvBEDG31dsFFZeTjVTG70zGSEAaBiB4WFMLx1EabGMlmYVZ58FBPwC+gcEUMrOsaKwD1ctaMGKRc2YV/sJWtua8EzrHXjn04tRViGisBAY2Kpi8ZJg0vPVY+W1Yfz8qUIEwwWaUkGhNtz6nSCoLf7czPLyANDXx14z00gvWJD96KlkyLUl6XjCIuAxxFTKXXGk63fQ0hLVkwYC0KRolZU05fL/+usL0NYmwO+P2jASwkjW7YbWrMAJy+mkqKmhmirC7Hh5OiQUio/IuVl6Ip/AZN9Xfb0aN0vO4aBQFJZKEQSKwkiKocjpg90O+EYoSovZ8v/b35ax510HKooHceU5O7F8wTacM/0A3jz8JWx+91bsOvglhBQJhLAc69AQRWEhTajL1bf81tUB11wn4ZLLHLjwChduJU5s3mzH8QPs2sarNKIwy8sD7HusrVXR2UniUg75IsRcW5KOJywCHkNM9tyVmXmMXimQzO9AkpjcKBjkAzrZjdrQkHzZ2NoqghDA72eNB6IYzfWGQiwac7nYtgkBGhuBcNi8VdmsVTgQYG5hPD1w4oQAmw04fpwtqxNV8hNh/nwFO3ZEbyvu51BU4IdDGEaJywNCmCaa2zGwhwBFw4xhfGlWC/7t5ldxTt2HePvTS/DyRzfiB9t/ho7OQhDCzHGg4xn+WTNd7jt7RPzbUy74QkXwhwpxrM+Ftz8m2ECMCowjRxwoLDRXaXAk0+5yyd3AAIEoIm1CHE06Lpv8/ESERcBjiMmYu0pmLpNquONoB3TqCdNmowiHmaMa76Lj9ox820zIHyWH2G2btQqzcT7sNe5LYbMxgk9VyTfDwYOiNksuHFJR4hqCyzYMl1NGKAQo3NFMAAQRKLD78OVzX8d1F2zDBWfuAe37AtqGV+Bf3/gx/CHWemZzRJsUSkooenqMbmUAk6O9s0fEK69IOHK8EFW1TnR0FmLQE9/rrY/gzVZlHg9w771OVFdTjRhTaXfdbqC2Vo2khOIbZWIxFdNx2cAi4DHEZMtd8ZvE42GRYTgMeDwEVVUqamqizRSxle4DBwSsW+fU0g5Rks5sQKeeHFQ1mgcWBIpQiPniciIoKWEFn3BYwMyZTH/75JMSbr+dPfCy78ajAAAgAElEQVTmz1fR3U20iJaTiCiy9AUnYSHyfOTEkm4ln6O9naCq3IuGuiEU2L04dYot0ymA2jo2r81hC+Cys97A1edtw8Vz/oiPOxZh9+Fr8D+f/ABLr3ah+guA/11j8wP3YWDtwCqGh0XIMjBntopv3CJCRhke+VE5gmEnAILuYeDwYUFzVzMeoxD3/7z7jq80bDagsjKaE9+wIahNADHT7no87Lfh9bLvIxWhTsV0XDawCHgMMdlyV5s22bWCCzeEUVVm1u1yqXA44oc78hv56FFBSzvoc6uZDOhM5K0QDketLSllN34oRHHFFWH88Ic2vPBC/CTijz4SIQgsN+x2M/Lo6CAa8fLCHs8fc2JJu5KvBkHCHiyeH8TJk9HrIcvsGAocQVx+9lu4ePlWLD7zDRw82YidH6/Aw7//R4iuEs1n4fDTTN+7fn188wPAdLodHXbMnV+MZdcQXNbkQOsuCffe60R/v1He53CYS8H0ETzPWfNrxRUa7GEL6F3cnnkmOgEkVrvb3R0lXz0SEepkT8flChYBjzEmQ+5KXzgLRexYBSFKUqoaXfr39xs/y18Hsks76JHIW0FV2THxMTuiyCLkp56ScPnl7MERa6CuKCwSbW9nx+F0UoPJOcC2U1PDXctYjri8nCYmXargj61ebPu9H6e7Qpg+TcXcuSpOnmS3lUDCuOzsd/Hlxm248pxX0TE0G6+1XY1HfvcPGA5WAYgY0HgBQNVIuLnZhh/8c1DXEEG0jrMLr2Cm5Hz0El+l8PPVP/DKyym6ugTENr7or//atSHs2hVNU+hN381c3GLztt/7Hotwly83n4aRiFAnYzouH7AI2IIBTz5px1NPSVrxiEe+oq4tXxAYcbjdFGvWyIbhjiMj0RbUdNMOiYoxibwVVJVAVaPkyyHLBM89B21wJQeXfwHQ0hinT7Ob3x5ZCfNzPXaMFeH4dh0OalxKUwqijICEPdjzVhDPPR1dSp/oEHDyJMU3rnkbdeoOnFf1Knq807DzTytwz0t3otdbp9k7VlVFx8erCtDVKUC0sVyv3y+AEhFULIyM3ylM2FjBl/Kxnsn9/WwSc22tgpoamjCCb2pSIlpedv246Y8oxru4JcvbZkqoky0dly9YBGzBUGg7flzQbkAeZeqjTe7QVV5OtaGRQPSm4WOEOBhxUgSD0Ioz+/YJOHiQkbbTSdHdTeB2syXvrl0iduyw4fzzFdx9t4yVK0OGgZR33inj4EHRoDTgkCSKI0dIhAyiTR/6/G5BAUV9PcXHH0fJipMzIdHlt90O1NRE86cvvhDG0kt6GPlStsHmLZxAKM6q3ofLZm/FpbN2wK+Uo/zcqyBXPody5xmorxHxcbMNYiTrUlXFot2+XopAkESaLQiUUAGCIy6c6HfhtT3m8rxY8AgzVmPNyfPuu2VD2sBsOvWCBZw8qUHjG+vilixvmymhTrZ0XL5gTUWO4PM6FVkf1bS3E20eGI8uQyFGYIQAxcXR/KJ+oq8+enU6KbZvt2kaWKeTYmgoWgziNzj/N1/ql5SoGBoy5nz1FXY9Vq4MaVG6HnV1Ki66SMSqVT5DDph7RdjtjKgCAWBwMKqCMI5sZ9MmHA6K2bNkFErDQGgYPm8Y06fT6Pici8L44f1HcOmZ23DprO3wBl3Ytm8FXj90DYbVM/HbzeYmQA895MCJDgFeL3Ciw4lBrwveYCH8cgEAAXY7m/W2cGHyJgb+Xa5b59R0x34/0byWq6oofvxjY8420eRlwNhwwvP4lZVG7+Xly12mLd6iCGzf7svY0yEVptI9aU1FtmCKWGkWz/UqCruxOBFXVVFNlqQf2x7bntrZKaCkhEY0uozQJV1BP9Honb4+tvTnkGWiyy9TQ2PDrl023HmnrDmr6ZUWt90GnH8+S3X88pcscgZY6qGsjGoPGH6ePJ/MYbcpKCnwoNg1jGklPjYhuF+AQAg6OihE72fo2rUVgn8rvnclQcvH1+DOf38an3TPA5fASRLFO3tE04kRV18r4fF/LcOhz9zw+iX4IzzN8+x2u7k/RiLodceCEFVx3HyzbCC/ZNErI/p4a8xY8kyVZpgM9Y2JBouAP+eIVRooCjH4DwBs2X7zzWzpr9d4JpprNjhIDD4GoVC0VTXR6B0ufeLgCgpVhWFpHQwS7N0r4u67ZTzzTCAu4lq2zIUXXhDR3GxnEev06LKaezIAbF/8QQOoKHGNoLTQg2nVHnhGBG35PTREUF9+BFeftxXXLNwKhz2AlgPX4DvP/QR/sXQu/u92SRvZo11HCXjkEQcqKijq6gjrPvsS88ldslTE3xERt99ujzRWRHOuAEtFADTtYpRed6x/GLW1idCnhlKpDtIhTytvm3tYBPw5QKJpEa2tIk6fJpp8yemMRqQ89ytJFFdfHUZzs11bmv7pTzY0N7OfTmFhNPoMBhnxxk7y5UTH92M2escW80vkCguvl2hRHYck0ThZlB6xDwZewDtxgi3xGUmpKLD74Pd4YKNeuN2qprX1eilq3R1Y3rAVV167DWWFfWg5cA3+14sPY/+JhQBY113pZ2GUllIMDrKONLudwiEBPQMF8MuFCIoFONbnxNsfQ+s+AxjZXXGFgqNHBUPOlZ8bwEgtnU6x9nYhbgoyf12PXKgOrLxt7mER8BSHsZssfugk86olkTZhlosNBIz5P70eWG8oAwDDwwQ+H0F5uQpZjha8YhEOs6i3uFiFx8OUDO3tjGjr6lQ4naxIZ2zcoPD747Wl5eWJR9K3tLBCXqzlpdsNFBVRzG3woVAahksagUAUoA5wFVBUVFIMnOrCRWfswNXnbUNtcSfeOnYVfrzjAbxz+AKAGKcz2O0Uhw8ztYU/YIOCAsDpwtGTRfAHbEwjrUQJMVYPy6PJWKXI/Pkq7r6buc+n0ymWLrHmKnq10gy5hUXA44ixsKbUR4PDw0BPDyPaw4cdui6pKAHYbCQusnz00ajOlJvecKgqI2Qu60oEbpQTCDAJ1uCggKEhgsFBlgv+u7+TTQ3QN26UcPCgEJfrdTop1q1zGq4dADz+eHSfXBNrE4KorRjG0ouGMNAfxlAfQa9MIEkEs+pO46GbmjGvaAeCfe1488iV+M8P7sGBzougUhu8XkBRmERMO2cqQIEDh48XoaCoAMd7IjranujAUf14HSA+Io2fnmGU6K1bZz4uOhGRc/BVysgI60bkvyk+kUSvKLnlFjnt39tUslGdSLAIeJwwVr3w+lbTri6eYwSCQaOFIF/CimJ8/359vYpDh9hPJbYKzgtYqmrU5MaCEJZj9vkIAgFmysNf9/kIfvITB7773aBJ5V+OXBdjt51Z2ysfVlleTnG6W4Hb6UGR0wO34EdJAcUlF4fR3GxDqasfVyzaieULtuKsukN4509NeMN3O37/9iWQnMYW4MJCoLxcxdCIE0MjhQhRFwrcTpw8LaC8nMLpgiEHKwgwbf81W+oniybb2wWNTPXRfDIiP3BA0BphiooQZ63Z3GxHZSXVho2ajYMyg+XbkD9YBDxOGKteeL5Eje0M42QZ26pqRhSsW0qMk30B8UboiUApK85xaZvmiYsoqf/61xLuuSexNzCPvrq6WNqDgxOVd4SirmoIZ9QM4swKX3R8DwH+9vZenN63Cz/6qx04q3of3j16GX7z1s1485PLAEHCtGkUvf1EN36dIBBywSu7UTe7AH+3xhid799PYxpOWM7a7yfo7hbQ3c3OixNnpkt9p5Pi0KHoE41H87W1ia07YzXYHNlaa3JYvg35g0XA44Sx6oXnS1Ru18hRUcF0t7FeDolGxtx5p4yNGyXo3cZikWwMEG/m0OePOQnzIlvsQ0K/f/2Nrm97HRlRMTLgQ0WBBzUuHxwOwDNI4awCZtd7sKj2dXxpznYsCryPPeLF2LLvBjxyeiOOHi+ELBOoCnMoY6buBMdOFqFmugt+uRAU7MBWr4kfoKknO30hzWaDoRApy6wwuW+fYKqbTrSUNxnjlhLJflOJCDid35vl25A/WAQ8ThirXngeQd57rxMDA8bClMulQpbT83C95x42Mv2DDwTTSBgwds1x8EaHZJEyj8Zj86aJcGZ9GN0nfSiUPBB9frhKIvsiFA57EJfNa8WKhVtx0ey38cHRC/CHT6/GT//4OD5rd2ttwAE/QVgBgiE7fHIhZKUQQcUFVSWoPVOBoJCk10SfezV7cPCctyRFfSoAJPVP1sPvJ6byMjOTeY5kvylKkfXvzfJtyB8sAh4njKWmsqlJwb/+awCPP+7SzMoBRDrazKVcZujuJhAEEkeygLFtWRCgaVwBQJaj0TGP7Ph79X4ON98sJ945VUDCIyCKF99aE8TTEQ+GUEiAZAvi0rm7cf2FW3HRmbtx4MR52L5/BX608weQabFmclNSQnH6tIATpwrRP1yEYV9hZLoE+7vdzszda2oonnkm+Th7fWrkk09s2oONK0X4OXLIMsHzz0ta/lUPs6U8J71YeVky0kv1m8r292bpf/MHi4DHCWOtqWxqUlBaCjz9tNFaEUCcmiDRMfDoS++MFptO4EY2ssysHAMBY2MHIYwInU5msM4kbKzRIzb/CzXMvBcUL4ji015eshggVMbRd9/DzIt24EvzWnGk92y81X41fv3BQ+j3lmDGDBWOIgH2yHEGQi4E4UbAVoT2U3YIAhBSjA8EWWYkfOBAektrs9xrf380StWnESSJYmCAmBKw2VJ+7VpmqRlbhDMjPX1ag0+2CATMIvjsfm+W/jd/sAh4HDHWmsply4Dzz4+qDDKtbjudTDPM7B/N98GNbVwupuE10wVzrau5r25IR7oxUSgNwzb0AaTeHVgu7ILSdCYOjyzHTT/9Lj7tqEVYAWwiQWmpgtvXU/zPy4U49FkxAiEXVMrC7MIipgeeOZPi008F+P3G6FxVCTo72Rw6TmKpJFf6CLG8nEZ0zkZVSHk51ZQfsUi1lE+WD479DrkyJGqUFMVofm+W/jc/sAj4c4x0qtv66CoQINpQTb+fTaTgqQee6+V+BmVlFCMj8emKM85Q8fLLMcSqhkAUD0sxqDEsRVXYhvdC6muB1PcaVKkWcuVVGF74n1AdddjxvB2fnZSgUgAgGPK50TlUhLf223D1KooPTZbOZ5+twucjqKlhqwFOcHw8kaIQtLUJqK+naUmuYiNEp1PVCl967fKaNbIhB8xhFtVu2mQ37XCLTVdMBIWCpRHOHhYBfw5g1ooMxHeMAayg9MknNqxb54wbuulwUJ2TGesEA6B1z+knHff2Es2+kqcnRBFRPwY1qOV040mXQhz5GI7eFkh9O6HaSiFXLsPwgmfx1r4z0fw0m/I7fZqK998X4ZNdGBgpxrC/CISIEEWKX/+a4sMPvTBbOgPQutD0E5VZhM+i4Fh1iBmhJSOeZM5g6TRDpKs8GG+FgqURHh0sAp7iMGtFfuABY1QYDBJ0dDBiFEVGtGZDN3nXHFNOMNJ2OKJtw5ywGhqYeTcfnMlAUSD5UVHkgeDvAVFjoj5KIXoPQYqQLgQH5IplGJ7/C6iuWQDYlN+nn2ZqgoEhB/YdKsaJrhIoavRnzD19+eDKxEtnRszHjxNQGp1vxh8qei9cIJ7QUhGP2X5bW0Wt/VuS2MOOqyNi89/pKg/GW6EwESLwyYy8EXBnZyfuu+8+9Pb2QhAErF69GrfccgsGBwfx3e9+FydPnsT06dPx5JNPoqSkJF+HMaWRztLP7Abhsim9iTfP1YpiVA5mNnTT7QZEkWr+r4mq493dBB99JMJp92sdaaKgYH6jYiDfj985Cu/hnVhYtQ12G0VP0TJUnfcTKK45ccnP37/ihCdQgiMninGyqyDOtQ2I1xYnAifI2HPgZkGxkrhYQsuGePSeGhzBICPh2I60dJUHo1Eo5CJ1MN4R+GRH3ghYFEXcf//9OOecczAyMoKvfvWruPTSS/Hiiy/i4osvxrp16/DLX/4Sv/zlL/EP//AP+TqMKYt0l35mNwJvymDEyrSmgUB8G63Z0E3A6P9qWh3/0ggc1Ief/CiEEY+iOYWVlFBctFjBzx7pxNyi7bhs9jacJXjxlu9qPPH6E/i09xwABOvXy5qXLhvNUwRqc+PdgxUYGgJOdgmm5MtBKUsrpFNIiz2H+fNZp11sKzEnNP28PKNxUOLrrf+bmWZYlkkccaerPMhWoZDu7ycVSY93BD7ZkTcCrq6uRnV1NQCgqKgIDQ0N6O7uxmuvvYb/+I//AAB85StfwU033WQRcBZINwIzu0GYAxoDL/S0t0f/zcEtIWM9CRobKa6/vgBtbWy758wP48H7evGlSz0spxtQcNliwP4PojbZd8Hs47jqnG2Yjh0omdaHN48uxz/+zw/wwdFFqKyEptUFWLNCGEX46S8q8ME+NwCCxkY2bePoUSFpxx2HqgJ79zKv3FR5ydh0QaL8rZ60zCY+8+udCHpPDT0kydzdLV3lQTYKhXQLsKlI2tIIjw5jkgPu6OhAW1sbFi5ciL6+Po2Yq6ur0R87VtdCWkh36Wd2g5h1nJm95nYDF10UxvbtNshy1DP4xRftIERFicsDt9ODoZM+/NuPFTghG6ZAXLLoJK6Y8Sqk3hYIgVN488hV+NX++3Gw6y+gUmbuDgBDQ8x/IRByYUQuxqF3CrH9TcGwVN+7V0RJiQq/nxh0yLHgagwzr4vNm6NG8smW3YkITU9a2Ux8TuSpUV6evgF7rpDO74enTGK1yHqStjTCo0PeCdjr9eKuu+7Chg0bUMTdSzJEWZkLNlsSq60cIdHcpomIefOATz+Nf33uXON5rF4NlJYCzz0HHDkCzJ0r4rbb2N/4aw0NSPjas8+KmD2b/U0Sg+g+5YWjxIcCyW8YNTQ8LGLnzgKsaOoBOncAp5qBkc+A2i8D59wLVCzBxtU2za5RAOBwAMNeJ3o9xRCK3ZpWNxhkUTdPASsKSyn09Iiw2Vh7bzBobHvmuV+mtogavPNJvwCbHvH444xEBQE4cYL9u7SUaaRT4eTJ6NSOsjJG8r29bB+Njex6LVtmPp6dfxeffQb86Efs+B0OoLISKC4WsX69iKoq86g0EdL9vba0AM8+G/1ev/nN9H4/Bw8yBz2AXVdZJujqAux247GuXs3+Y/sS8eyzdjzxRHRf6Vzb0Z7jZEVeCTgUCuGuu+7Cddddh2WRb6GiogKnT59GdXU1Tp8+jfLy8pTbGRjwpXzPaDHZBgDecIN5AWzVqiB6eozRx/nns/9iz/H88+O3a3hNDePUMYoypxcOmx8CUTBiExBUjP4Opa4BXHlOC1advRVq60GEyi+DXHMTQmddDAiRG3UwgJoaNpBSUW3wysXwqMU40edkueYQBcCO224HAgEWielHyutrctXVUT0ya3qgcDqBqioBPT3R0UcOB9Xar0dGiCY10+Ppp1VDg0oiTJ9udBtzuYCZM5nq46c/ZZ/v6Um+jW99C5g9Oz7Fcf75SsrP6pHu7zU2jXDoEHDffWyw6aFD8YSv//14vYWafakeXi9FT4837X0NDmYnSZts92QyjPlQTkopHnzwQTQ0NOA2Hl4BWLp0KV5++WWsW7cOL7/8Mr785S/n6xCmNEaz9DMrrABsydl1Moi5szxYfcMQLlnsw4K5jDQ5JIkiFCJwO4exfOFOXL1gG86buRdvf/pFvN21FtNXXQiI8WbilIhYdm0hHv5RJYLhAgDQvHRra6mhdXbTJjv6+tiyV99JRwibT1dezoqDDQ0U9fWKQdtrswGKEk0P6FMrZuQLpF+xn4hTJVIVyRLletvaRGzYkPz3wzsfY5HoOlqStMyRt7H077//Pr7+9a9j3rx5ECKaoHvvvRfnnXce7rnnHnR2dqKurg4bN25EaWlp0m1ZY+lzg48+cuP73w9j715Ry+eVFCtQgl4U2L2oLPOxMT0RrF/PzHG49tZp82JhbSsuqd+GCxvexZ4jF2P7/hX4w6ErUFTqxN//vRw3CZiKhaA2N6hYBBCS1ujy1lZRGysf0AWm3GvC5WLNHm+95Yv73EsvuXD4sKKRRCyxm1XsGxqSj4CP3cdEyHdWVbnxwgu+hKPm+TGlGiWfDPqR93pHtoULza/XaPZlhql0TyaKgPNGwLmERcCjR2uriMcfd+Gzz1RACcLl8MEleTFzulcb1T5tmvGnMGOGih/84yCOvfc2yKkWzCt/G0cHz8cJdTleeONKfPxnpt+e3aDi5ltCUemYIEVItxgQsltktbaK2LhRwnvviVoUzI1+AJZauPNONqlZH/2tXu1CT48nLjKcP19Ba6vN8PDhygUz34R0jm8822+rqtxYtSqU8oGSyKQ9nYdOIp13ous1mn2ZYSrdkxYBp8BU+rLjoIZx99+q6Dst4/hnXgCKVgxzSBQqZbpgXom3CTL+YsZuXDZ7O7447w0oRecgWLkMofImULuxaeadPSKat9jQcdKGsupCXLvKhcuXZlZMSobWVhG33+6MUw5wT4r6euPP9/HHRQwOGiND3vxQV8fOj0d0ixYpiU2BUhxTJsSUD1RVufGFLygpI87RHmsmEX+ur8tUuifHPAdsYRxBKaD6QRS/5rXg6y8AlUUAVLtpVRUIBAmcDgqHI4QvnPEWLpu9HRfNfB3tA/Pwcf/VGDr/HlDJvFD6zh4RT/6sFCNyCXxyEdr7CfY+Aqgkd0TU1KSgooKirw+GZTAn0dgmkeeeAwIB4wOANz/09xPU11PN4KamhmZ0nDzq3bWLheGxTRhjnevkGu9Yqdj8+VFWHq1MLJN8tSVJyxwWAU8VqHLUwlH1g1BjaDR9mop33xMh6jS0AlGwZO67uPb8rfjyOa+iyzMTuz9bgf947270+2qwfr0MKkVvHh7ttnc4UF5dhGMnSzE8Eh/tZkJE6SzlFyzgzSRRsu3sJIaGEo4jR5hJux48eo4l7EzaZfXRHd9ebBPGWLffcs/g2Nbmri6Wa9drdceKBC3bysxgEXAekdc8IVWYSbniA1F8IDSc9O0rrw3jj2/aIdpUfOHMD7F8wVYsO68FPcPV+OPRa/Bh0X/if/4wE6dOCZg2TcXqm40FtXf22PHExgp45WIEwwXoGGDGNbETgD0e5rK2fLkr5Tmn2w5rpj7QO7jp0dAABALG7j/u7xBrsJNJ84O+ws+3BxibMMa6maKpSUFtLUV/PzWsDtxuS3kwWWARcJ6QF5s+JQCiRgg31qw8Ah6lnjwlwOGgIAACQeDSxn34p9UtuLh+GwZ9Jdi2bwX+5le/gZfMxNlnq/irJUEsWhLvGE7FAlDRjV9trkK/z/hzcTiogYB4rtXhYGmOVOecTqcVYL60XblSMfXWve02YHDQSNi8ay2WsDORj+mjW30XnD6qHo/2W7+fRPLgyd3bLExMWAScJ+REE6nKUbJV/SA0+ef0do1eL4Xk/zOWLdiGaxZtg0rt2PmnFfj2//k/ONY3x/C5lSuN0TMlNux+pxT/+dsKfHbMifp6FQcOCIhtZCwvp+jqiqYG9C5r6ZzzgQNC3PK5s5PEOZnFriRivXV7e9lnSkspnn0WuOEGGDSuCxeyz7S1iVnnJvWeGtyWs7+fFS8bGsYv12mZ4UxuWAScJ2Rl00fVaEpB9cV75qZA8xYbZpR9gssatuPimdsgCmFsP3AN/r9NP8WIeBY6O20IyarWTMEdyj77TABAQG1FUG1utP6hGA9siM4jO3qUjVavqTEWndxuoLZWQU1N1EwmNiURe856Mj11ig2wFGO6zPWTfxOtJPbtC2neuoLA3j84SPDhh8ChQw5s2BA0kT5lH6HGpkG4idFYKh/SOS6OsbKjtDA6WAScJ6QdmahBbegkUQKIXUqmA8HfDqm3BfcteRUu+wjePLIcD/zX4/i4YwGAqMRMlgGAYNo0VdtPSHHg409LoLhEgDAm3LhRiotMFYVNRXa7jcevl3El0oE6nRTr1jlx4ICAvj6i5Sn1bcZ6EtZ3WiVaSfAJw7H2jr29rD041znQiVrhz7cdpYX8wiLgPCFhZPK1IBAeiRCuN2XxLBGEwKnI9IgWCHIv5Ior8d+H/hf+8PEXQCHg1GkC3kZqtzNCkyTWIk6pgBG5GCPBEoQUBxoaVIBEo8WDB+NJVBQBQigaGtSEN7rZJF+nk8LlIvB6idZezNUDLheFz0c0Ix1eRFqwIEry+uhZny+WZWMxjMPnA9rbo2OVchnVZVvhz3ekmS87Sgv5h0XAeQKPTH67WUTXySDmzPLixlXDuGSxF0gwHTcVSPA0pL5X4ejdASHQAbliKXxn3oNw8V8ARETjF0W88TGf00a1sTwlJYyAnUUuDHmLcGKwCPoe/3SLR3Y7tGU9J5VHH3UY/CS0Y41s3uMR4HIxQtWTZX8/iUwKZhH63LlR0tUfj17rqo/KBYH9WxAoVDU61UNRoJH/RIjqJmqkaU2ymBiwCDjXoBRQAyCKD19e4sOVFwWRTVqBg8h9kPpeg9TXAtH3GUJll8M/43aESi6MOo1FwGRjsmaCXlWpQKE29HvK4Sh346EHXBgc9GHzZor2dpJwudrYqGLv3nj7z8ZGRpKJSMXlonGTfA8fJppSQh+xyjLRJnKwGXNIGFU//LAjLtVQUUG1FmoORYnaRfJCoMcD3HuvE9XVdFzynPmKNFtagF/8wpl1VG0V7yYGLALOBdRgjFphdD9iEhqE1P86pN6dEEcOIlR2GQLTbkaodAkgSEk/u2SxgiWLVVCxEKq9GBD5qAkFVVVAT0/q5WpTUxj794uaEY7TCdTUqLj7bmbOk4hUDh0SMHOm8WGjH2ukl29xTa7bDWzYEEg58n39eqchTeF2M1OegQGCigo2lbmnh8DlAsrKWCGQR82EAJWV+YuIk6UY8hFpMl8PIBxm28jmvKxJFhMDFgFngzTkYXo97vRpKlZeG45zCtODhD2w978BqbcFNs9ehEuXIFD7VwiVXmpq72gGZoJTDGor1gpqmYJP7q2pUbV2X7udGiLTTMiDjzUCjJ6AOaUAACAASURBVPKtykqatnyrqUnB5ZcrcRGb2w0sXKgYjGdOnLAjHEmr86g5tgEjl3nOVCmGfESauYiqJ2pR8fMGi4DTQYbyML0eFwBOdAiRf8fYNSo+SP1/gNTXAtvQ+wiXXAC5agVGznoUEBNPVtDv55VXHDh8rAQVNUX4q6+Joy4SnT5NtFHz+lRCW5sILuNKRCqNjSq8XmNawO0G1qyRNQ0u1+RmepzpRGxr14a0iRdANOccq0vOZZ4zFRnmI9JsbxdMpz5nel5W2/D4wyLgRFBDIMqIruss/Txu8xbzy9rcbMOSC7ywD74JR+8O2AbfQdi9CHLlVfDO+SdQW/rjV97cU4RHflwNn8wKan0jwMMPA5kur2MjOBY1MpUC/3cwSHD8eNRfIBGp8BSFeVQ1uqVtOhFbU5OC0lI24aK9XUB5OdUeJnqkij4zUS2kSjHkI9Ksr1dx4kT8CsfK304+fK4JWH+jzZtHceOqAC7/4kjEQUzOersnTxlvSpsgY9EZb+Hy2VtR+v4uKEWNkCuXwduwAdSe3IxeD0pE5rNrK8azm0vgk+Nv/kyX1zyC4xIvOXLaJ08aoyxKEZNnTEwq+Yqq0onYli2DNl4okT1iY6OCdevMC1iZqhbSSTHkOtKMjfQ5rPzt5MPnloBbXxfwxOMUTvsgKl0+BHtlPP90GAWxaYIsMH2ailMnFZw3fQ8ua9iGi+pfx4mBOfh44GoMnX8XqFSR0fao6GK53chUCSB3xZ32dqPESxRZc0QwyHTDvEGCL+M5wU+G5avZg6Kx0eghEUuwsQ8kLmnbuFFCU1O8/8Z4FLNiI30rfzt58fkiYJ1aofUVFbXF0T8RMKZpbrZlT8BUgW34Q3xvxU6U+F5H1/AM7D6yAr95/y5Te0c9Yot2K64FFn+xkBXUhPhoJ1fFnfp6VfO3BRjhqiqbQizLTPtbUTF+lovJYJYq4NN5OWIfFOvWmRc0+YMl9oEEsFzy3r2iweJRv/1czuZLl0T1kb6FyYupTcBqKFI487Mimk6t0NlZYPqRU6cyJBiqwubZH+lKexWqVAnbzOXYc+I3+O836hPaO+oRLdoR+EOF+OBQCd48UJjUZyBV5JUOOfHt7NgR/RnwZgZBYGkHmw0YGmLNFG73xMkz6lMF3AJzxw4bnnoKuPPOxMXIVCuH2AcShyTFu7RxZLMamKgNGhbGFlOLgNUwiOqL2jYmUStMn6Yapv1yMJ+EFKAU4shBSH0tkHp3gopFkCuvwvC5/xtqQT0AYOF0YKGJvaMZXv69C4P+UniDxVBo9CtJlc91uSgOHWLn0Nioar4MiW7u0tL4UfRNTQoWLVLQ1iZAlllbsD0ScOsnEvNminznGdONCvWpAn20un9/bK7aiFQrh9gHEkd5OR1T9YSFzwcmNwFzeZjqj0S66RfOVl4bNkjFtNdXJvBmoBSi7xNIvTsg9e4EiAi5chlG5v8Mimt2wv0k1gMTqJGC2vttFaazvRLd8HxqsD5H2dUVlX/pb259LvPb3wZ+9KP46PCee+QIaVEcPqz3vWUNDrLM2oXz7fyVSVTIr01sh1ww8sxLRGSpVg5NTQrOP1/BwYNCnMl5LqN/qxXYAjDZCNgw68wHomafA4tt260/E7jqqvg0geA7AkfE9AZqCHLlVRg563EohWdFDQ8SwEwP/NOfFyMIJ77Y5AJIdNmbST7XzK2ss5NohSJ+E8dGh7295tGhPo95/DiLgp1OikAgSvCNjWpeu8fmz1fw/PMS+vujxuw872xGpvyaxZrxOCLcmojI0snZ3n139IGkRy6jf6sV2AIwWaYid3ZCCA+zETyj8FVIhvLyIvT3jwAABP+JSHphB4TwMOSKZQhWXgWl6NyUpKvHQw85cKJDgEpFeIPFGJGLNfcxvVdtptNk580rjCMegJHmn//s1Wwh29uJ4X1OJ8HMmYq2f7PlPgBtzhjPB6sqU0Q0NKhwOjFqT4XYCF4QzG0pubewfsqvfhsPP+yIO8czziBQFAXBIBmV/0Mm04CzwWgnCE+licGJMJXOcVKPpe/r+AQknN8votw5BN9nv4PU2wJB7oZccSXkymUIuxdqkWqm+OubKzAcKNGaJTgSEUq6N3wqAuY3tz6dADBycrkUeL1AdTXFRx+JcdHmhg1BPPmkhP37Rfj97HlDSLQwV18f9VmoraWRkTgsgj14UEyron/99QUGsx9ZZiQvCIzoORwOivr6aMuy2cNi40YJe/eKWqpAFEV0dNA4Y/jxNk43w2hIfiqRUyJMpXPMmoC/+tWvgpDY9lI3Fi1ahG9961soLCxM8MncIV8ETOQeSL07IfXthD1wHIHyJsgVyxAu+QJAssvOUGLT/BjWrXebLjNjI+BMcf31BfjgA1GLTgWBkWRJCcXMmdSwpB8YiOYxy8pEDAwoWt5UT+KcsBoaVBw7JuDoUaJzLmP7IQQoLuamOoJGkDzVkS7pxT5AuOkPIdEUAv/33LkqVq4Mmc5/49vXE1lvrwibTYnrfhvtNZ9omErklAhT6RwTEXDK0O7iiy9GXV0d7rjjDtxxxx2YNm0azj33XHR3d+P73/9+ro8z7yByPxxd/wX3x+tQsnc1bN4/I3DGt4Gr/gjf7P8f4dKLsiJfKhZCcU6D6prFGi0Ee5xHLsdoc4lNTdFCISEsNa4oTL3Ah2E2N9txyy0y5s5VUV8fjXD1Prx6cFLm0Zj+77xASAizkeTvZS5n0c/GFsQ2bzav9MeCS94ojWqQVZUpDzZsCOLgQXNjIb79piZmyLN9uw81NYgjX35eFixMNKRkmvfeew+//e1vtX83NTXh1ltvxfPPP48VK1bk9eByBRIagr2/FY7eFogjf0Ko7FIEpv01QqUXA0Ik5BLsyNQpPRrtlgBC/KXMl+PUwYMizjgj6lYWCjFy5JEkVz489ZSExkY18jeCuXOBwUGKoiKgv59No+BRtCwTeDxUM8vZtUvUSFgQ2HtEkelh+evcZUzv8avP0ScivfnzVXz0UZRU+QKLELYv3gJ9881MVvfoo/G50kTbb2gADh2Kf69V3LIwEZEyLBgYGEAwGCUmWZbR3d0NQgiczvRsEscDJOyBdHoLig7ehZIPr4M0+BYCNTdg8ILt8M57GKHyK6LkmyGoWITWPTPwrXsasey6GVi3vgitreZRGo/Ovve9ICgFHn3UgXXrnAnfnw7a24WILIpi7lwVNhsjR0aiLB3AxvYQ+HxsHND3vhfEpk3Qxv04nRShEAzyt85OAY2NrKngzjtlNtaesPfa7Wwf3OAGiLYn838TwsYBHT7MCoD62W563H23jLo6Vds+Iazho6CApyFYDpc5sCUmT7PXv/lN82tmtupobRWxbp0Ty5e7Rv2dWLCQDcTvp8gj9Pb24oknnsDQ0BA+/PBDPPbYY1i6dCnOO+887NmzZ0yiYP9wf3oaX8UPqf91FBz/BVxHfwSiBiBXXQ3v7AchV62E6mowbesFgIICCX5/4n1QYgO1l0J11OL1P5bjh4+4MTjIJFuDgwS7d9swY4aKWbPiSYcXxfTv37rVhldeseHf/13C7t0iioqo6WfNsHu3iMHB6HJ/ZARQFCbf8vsJFIX9zeGgKI14/Zw8KeBrXxMhCEHs3m1DVxdBOMyOh72XNaGIIsF114WxZImKc89lOmCbjWDmTBV1dRQ2G0FdHSNOvtQXRWBoiEBVoY0HkmVmkP7b39rx7rvG85s1i2L2bGZdabezOW61tRR1dRQVFeyYHQ5geJjgG98IoaiIYvfu+BXGHXfIcdfsvPMcKC8P4ORJAcPDBLNmqbjjDjlu1WH2nST7DicaCgsd8PmyN4yaDJhK51hYaB7spaWCeP311/Huu++CUorFixdj6dKlOT/AZEhahFMCsA++Bam3BfbBtxB2nwe5YhlCFVcwH4U0oZeh6UHFopjJEomn/yYq9MS+n0ep+iJWfz9BRQVFdTUjN64uMFMTxEqY9EUwve5XXxQTReCDD0T09Hjw5JN2PPaYQxuGyfPINhvrrnv66cQTKjiefNKuFfnKyigEgcLj4U0bbOy9KEaVDEDiolw61zNdxUC6hZtMv8OJhqlUoEqEqXSOiYpwKXPA77//PpYuXTrmpJsUagj2wbeZeqF/N5TCsyBXLoOv4Xug9rJRbz5VbjfTLqbY1/VFLI+HaqTZ1UVw6hR7T12dmrATjOeWN26U0NYmgFJg1iwVpaUUAwMUlMIgLQNYGmHNGuDwYRdOn2bpAVUlhtHw4bCZ7WQ8+NSMykqKyko++42rINi8OZ7a0OeFs+1O4+ecSxmZ1YlmYSIgJQE//PDDGBkZwapVq7Bq1SrU1taOxXHFg4ZhG3ofjt4W2Pt3QSk4k5Fu/V2gUlUOdkBMo10zZNrFFPt+fRFLrxwIBKI6WO69ACQmLq+XGGaweb0Ed9whx0m2PB7A4yEIBlnOt7+faD4Per8HRQF8PoL2diS0XwTMfQwcDqods15BoR8HNJrutFzD6kSzMBGQ8nH/4osv4uc//zmGh4dx44034pvf/Ca2bNkyFsemwfXpD1D63tVwHX8KSsEsDC/8v/AseBbBujWjJl8q2KFKFUDRLKjOupTkCyBjeVns+/VFLLOGCiAq8eJOX7GFokRmLm1tIjZsCKKhQYUosiV1ba0xGnY4KEQRsNuj5EgpUx+IotF+0QxmRFpeTrVj5ufHX+eor1cTFr70UrJnnkmdAhkt8iURtGAhE2TUCSfLMv7lX/4F//Vf/4W2trZ8HpcBI+/9AGH3QqjOM3K0RQJqK4JqK9Zmr2Wab8q0i0n/fqeTmee43TC00nK5FwCtQ02fK+bYsCGIRx91mBr4mHXZLV/ugqoCNpuIcFgxTAuWJJa7VVVoSge+f06KsUiUPy0spKipoThwQEBvL4lLg6RqqMgFMvke891unE9MpfxoIkylc8w6BwwAhw8fxksvvYQtW7Zgzpw5eOyxx3J6cKkQrFuTk064XEwN5sg0Jxn7fn7ze70CenujkSLPB7NpwuZDJTdvtme0hI59L59OLMssF+zxEAP58n0mShkkmwmnH+0TS24TzYJxMkz1sDC1kZKAV61aBZ/Ph+uvvx4vvPAC6urqxuK4coj4aHciQH/z68mqtpa9FggQDAwgrr0XYCmA730vmPYoHDPCdLuBDRvYUv8rXynIyH4x3QGZseSWSUPFZMFoplpYsJCSgB988EFccMEFY3EsOQUV7JE5asWmSoaJhESRWKKlfn29mlHhir/3pZdcOHwYce/Nxn4xm+hxNIWviUh01lQLC6NFSma64IIL4PF4cPToUUNH3IUXXpjXA8sWVCyEai9Jq5g20ZGOeXi6N3pTk4LVq4GeHp/p37JRIWRKitkOsJyoRDfRUioWJh9SEvDWrVvx2GOPYXh4GNXV1fh/7d17UJTX/Qbw52URNagoyKJ1CN6lVAmm9YI34hogLRJARU1itNTWxmiIYewIZrQNZvCuVTLemjFa6+i0UUiqY2KlxguJMh0vaAYTfyogGdlFXREvLOzu+f1BeKPCstx2z7I8n7/cw7Lne+Z9fXj3vZxTXFyM4OBgZGVlOaO+RqlZrt3H5gKWbZUzb89q6hFtc0KxueNx1aDjvcTUUnYDeNu2bTh48CDmzp2L7Oxs5Obm4ujRo86ozS7h0Qmigw+EpmuTJkpvS1z1QlFzQ7E543HVoOO9xNRSdvdgT09P+Pn5wfLjHftjx47Fd99916gPT0tLQ3h4OCZPnqy2ZWZmYvz48YiLi0NcXBxOnDjRpIKF4gGrpw8snZ+HtXPgj3c0uGf4ujJnhmJ9gVZRAej1itSJdHgvMbWU3SNgLy8vCCEQFBSEPXv2oE+fPjAajY368ClTpmDWrFlYsmTJU+2//e1vMXfu3CYVWnML2Y+nGZq5QgU1TmPO7Trz6O/Zc8dPzn1RO/9x7emP6dNbvXubZDzBR+7FbgC/++67ePDgARYvXoy//OUvqKiowJ///OdGffiIESNQUlLS4iKtHXq6/J0MrqA17hRo7Lnd5l5Qa45ng85kUuq9PW///g5ODeDa2hi41FwtXhNuy5YtePvtt23+vKSkBG+99Zb6+HJmZiaysrLg7e2NoUOHIjU1FT4+Pg32YTZb4OnZNuZqPXoU2LkTuH69ZnLw3/0OiIpyTr9Ll9Ztz8hoWv8zZwL/93912wcNAvbtq9vnJ5/8NNakJOeMdeRI2HwK8OxZx/dP1FpaHMAJCQkN3hHxbADfvn0bPXr0gKIo2LRpEwwGA1auXNlgHy19HLExR4at8dhjQyvdAnDofayNmV6xMWOsfWz5WfU94ixLQ2M9eLCD2zy+2hB3ekzXFncaY7PXhLOnqfnds2dPaDQaeHh4IDExEZcuXWppCQ2qDcUbNzyeOl/oiIs2tu4M2LTJy+E1tNZFsaasPiELL36Ru2hxAD+7YrI9BoNB/fexY8cwaNCglpbQoIZul2pttsKuoKD+9tasobWCsy2E28SJljozvrnisvNE9jj0ylZKSgry8vJgNBoxYcIEvPPOO8jLy8OVH1dN7NOnD9LT0x1ZgtNvl6rvq7EtrVlDa10UaytX9nnxi9xBiwO4oVMQGzZsqNOWmJjY0i6bRObtUrWCg6149KjuN4XWrKE1g/PZiYL27euAVas6uswcDETuolEBfPfuXVy8eBGKoiA0NBS+vr7qz3bu3Omw4lqDzNulakMQgFNqaO2jwoZuSQMce1GRqD2wG8BHjx7FsmXLMHToUFitVly5cgUrVqzAyy+/DABPhbErcvZXatsh6Ppf65/V0EXFhw9/OqJ3lclxiNoauwG8ceNG7N+/H/369QMAFBYWYv78+WoAtwWucL7QFWpoqoYuKj65Fl0t2ZPjELU1dq8C+fj4qOELAH379kX37t0dWhS5BlvnqG2d9pc9OQ5RW2P3f8y4ceOwdetWlJWVwWAwYNu2bYiMjMTjx4/x+HH9q+aSe7B1S1pIiOvfK0zUFth9Ei44ONj2LyuKUxbndMbTMO701I0tzRljfWu7AfVfVHSFe3Hbw3YE2sc43WmMzV6Us/aeXWqf3OmiIpGrafR9wFVVVeqcwADQuXNnhxREbUNbvKhI5GrsBvAXX3yBlStXQq/XQ1EUCCGcduqBiMid2Q3gtWvXIjMzE0OHDoWHB69yExG1FrsB7O/vj9DQUGfUQkTUrtgN4DfffBN//etfERkZiY4df7ryPXDgQIcWRq6hNVbZIKL62Q1gvV6PXbt2ITs7Wz0FoSgKcnJyHF4cydWcpeeJqPHsBvCePXtw9OhRaLVaZ9RDLqS5S88TUePYvar2s5/9jOHbTjlzLmWi9sjuEXBoaChSUlLwyiuvPHUOOCIiwqGFkXzOnEuZqD2yG8CXL18GUHMqopaiKAzgdsCZcykTtUeNOgdM7VNbWZ6IqK2yG8BCCHz66acoKirC4sWLUVJSAoPBgBdffNEZ9ZFkfOSYyHHsXk1ZuXIlzpw5g2PHjgEAvL29kZGR4fDCiIjcnd0APnv2LNatW4dOnToBAHr06AGTyeTwwoiI3J3dAO7YsSMU5af1v6xWXgEnImoNds8BDx48GJ9//jmEECgpKcGOHTvwy1/+0hm1ERG5NbtHwAsWLEBeXh7Kysowffp0WK1WzJ8/3xm1ERG5NbtHwG+//TaysrKeaktISKjTRkRETWMzgM1mM6qrq2G1WlFZWalOxH7//n0uxklE1ApsBvC2bdvw0UcfQVEUhIWFqe1dunRBUlKSU4ojInJnNgN44cKFWLhwIdLT07F8+XJn1kRE1C7YvQjH8CUicgzOK0hEJAkDmIhIEgYwEZEkDGAiIkkYwEREkjCAiYgkYQATEUnCACYikoQBTEQkCQOYiEgShwZwWloawsPDMXnyZLXt3r17SEpKQlRUFJKSklBeXu7IEoiIXJZDA3jKlCn4+OOPn2rbsWMHwsPDcfToUYSHh2PHjh2OLIGIyGU5NIBHjBgBHx+fp9pycnIQHx8PAIiPj1dXWyYiam/srojR2u7cuQOtVgsA0Gq1uHv3rt3f6dHjOXh6ahxdGvz9uzq8D9k4RvfRHsbp7mN0egA3h9H4yOF9+Pt3RVlZhcP7kYljdB/tYZzuNEZbf0icfheEn58fDAYDAMBgMMDX19fZJRARuQSnB7BOp0N2djYAIDs7G5MmTXJ2CURELsGhAZySkoKZM2fixo0bmDBhAv71r39h3rx5yM3NRVRUFHJzczFv3jxHlkBE5LIceg54w4YN9bbv3r3bkd0SEbUJfBKOiEgSBjARkSQMYCIiSRjARESSMICJiCRhABMRScIAJiKShAFMRCQJA5iISBIGMBGRJAxgIiJJGMBERJIwgImIJGEAExFJwgAmIpKEAUxEJAkDmIhIEgYwEZEkDGAiIkkYwEREkjCAiYgkYQATEUnCACYikoQBTEQkCQOYiEgSBjARkSQMYCIiSRjARESSMICJiCRhABMRScIAJiKShAFMRCQJA5iISBIGMBGRJAxgIiJJGMBERJIwgImIJGEAExFJwgAmIpLEU1bHOp0O3t7e8PDwgEajwcGDB2WVQkQkhbQABoDdu3fD19dXZglERNLwFAQRkSSKEELI6Fin08HHxweKomDGjBmYMWOGzfeazRZ4emqcWB0RkeNJOwWxb98+BAQE4M6dO0hKSkL//v0xYsSIet9rND5yeD3+/l1RVlbh8H5k4hjdR3sYpzuN0d+/a73t0k5BBAQEAAD8/PwQGRmJ/Px8WaUQEUkhJYAfPXqEBw8eqP/Ozc3FoEGDZJRCRCSNlFMQd+7cwYIFCwAAFosFkydPxoQJE2SUQkQkjZQADgwMxOeffy6jayIil8Hb0IiIJGEAExFJwgAmIpKEAUxEJAkDmIhIEgYwEZEkDGAiIkkYwEREkjCAiYgkYQATEUnCACYikoQBTEQkCQOYiEgSBjARkSQMYCIiSRjARESSMICJiCRhABMRScIAJiKShAFMRCQJA5iISBIGMBGRJAxgIiJJGMBERJIwgImIJGEAExFJwgAmIpKEAUxEJAkDmIhIEgYwEZEkDGAiIkkYwEREkjCAiYgkYQATEUnCACYikoQBTEQkCQOYiEgSBjARkSQMYCIiSaQF8MmTJxEdHY3IyEjs2LFDVhnt0vHjGsyb1wnR0c9h3rxOOH5cI7skagO437Q+TxmdWiwWpKen45NPPkFAQACmTZsGnU6HgQMHyiinXTl+XIOMjI7q6xs3PH58bcLEiRZ5hZFL437jGFKOgPPz8xEUFITAwEB4eXkhJiYGOTk5Mkppd/bt61Bv+/799bcTAdxvHEXKEbBer0evXr3U1wEBAcjPz7f5/h49noOnp+O/7vj7d3V4H7L98EMHeNaz1X/4QQN/f/f4z9QetiPg3HH+8AOk7Dfuvi2lBLAQok6boig23280PnJkOQBqNnRZWYXD+5HJ378r+vSpxo0bdb/4PP+8FWVllRKqal3tYTsCzh9nnz6dnL7fuNO2tPWHRMopiF69eqG0tFR9rdfrodVqZZTS7rz2WnW97TNn1t9OBHC/cRQpR8DDhg1DYWEhbt68iYCAABw+fBjr16+XUUq7U3PBxIT9+zugqMgDQUFWzJxZzQsp1CDuN44hJYA9PT2xfPly/P73v4fFYsHUqVMxaNAgGaW0SxMnWvgfh5qM+03rkxLAABAREYGIiAhZ3RMRSccn4YiIJGEAExFJwgAmIpKEAUxEJAkDmIhIEgYwEZEkDGAiIkkYwEREkiiivplxiIjI4XgETEQkCQOYiEgSBjARkSQMYCIiSRjARESSMICJiCRhABMRSeJWAZyWlobw8HBMnjxZbcvMzMT48eMRFxeHuLg4nDhxQv3Z9u3bERkZiejoaJw6dUptP3nyJKKjoxEZGYkdO3ao7Tdv3kRiYiKioqKwaNEiVFVVOWdgT7h16xbefPNN/PrXv0ZMTAx2794NALh37x6SkpIQFRWFpKQklJeXA6hZAPXDDz9EZGQkYmNj8e2336qflZWVhaioKERFRSErK0ttv3z5MmJjYxEZGYkPP/yw3kVUZYzR3balyWTCtGnT8OqrryImJgabN29usLaqqiosWrQIkZGRSExMRElJifpZTR2/s9gaY2pqKnQ6nbotCwoKALTN/bVFhBvJy8sTly9fFjExMWrb5s2bxccff1znvVevXhWxsbHCZDKJ4uJiMWnSJGE2m4XZbBaTJk0SxcXFwmQyidjYWHH16lUhhBDJycni0KFDQgghli1bJvbu3eucgT1Br9eLy5cvCyGEqKioEFFRUeLq1ati9erVYvv27UIIIbZv3y7WrFkjhBDiq6++EnPnzhVWq1WcP39eTJs2TQghhNFoFDqdThiNRnHv3j2h0+nEvXv3hBBCTJ06VZw7d05YrVYxd+5c8dVXX7nEGN1tW1qtVvHgwQMhhBBVVVVi2rRp4vz58zZr+8c//iGWLVsmhBDi0KFD4t133xVCNG/8sse4ZMkSceTIkTrvb4v7a0u41RHwiBEj4OPj06j35uTkICYmBl5eXggMDERQUBDy8/ORn5+PoKAgBAYGwsvLCzExMcjJyYEQAmfOnEF0dDQAICEhATk5OY4cTr20Wi1+8YtfAAC6dOmC/v37Q6/XIycnB/Hx8QCA+Ph4HDt2TB1nfHw8FEVBWFgY7t+/D4PBgNOnT2Ps2LHo3r07fHx8MHbsWJw6dQoGgwEPHjzA8OHDoSgK4uPjnT5OW2O0pa1uS0VR4O3tDQAwm80wm81QFMVmbf/973+RkJAAAIiOjsY333wDIUSTx+8KY7SlLe6vLeFWAWzL3r17ERsbi7S0NPWruV6vR69evdT3BAQEQK/X22w3Go3o1q0bPD1rltHr1atXg6HgDCUlJSgoKMALL7yAO3fuQKvVAqgJsLt37wKoO87auhs7ftnjfHKMgPttS4vFgri4OIwZMwZjxoxBYGCgzdr0ej169+4NMWl9qAAACSpJREFUoGZh265du8JoNDZ5/M727Bhrt+XGjRsRGxuLjIwM9TRLW99fm8rtA/i1117Df/7zH3z22WfQarVYtWoVANR7nkhRFJvt9WnoL7mjPXz4EMnJyVi6dCm6dOli831NHWdTxu9oz47RHbelRqPBZ599hhMnTiA/Px/Xr1+v857a2trqtnx2jN9//z1SUlLwxRdf4MCBAygvL1fPT7fVMTaX2wdwz549odFo4OHhgcTERFy6dAlAzV/K0tJS9X16vR5ardZme48ePXD//n2YzWYAQGlpqXrE6WzV1dVITk5GbGwsoqKiAAB+fn4wGAwAAIPBAF9fXwB1x1lbd2PHL2uc9Y3RHbdlrW7dumHUqFG4cOGCzdp69eqFW7duAaj5Ol9RUYHu3bs3efyy1I7x1KlT0Gq1UBQFXl5emDJlis1t2Vb21+Zy+wCuDSUAOHbsGAYNGgQA0Ol0OHz4MKqqqnDz5k0UFhYiNDQUw4YNQ2FhIW7evImqqiocPnwYOp0OiqJg1KhR+PLLLwHUXJHV6XROH48QAu+//z769++PpKQktV2n0yE7OxsAkJ2djUmTJj3VLoTAhQsX0LVrV2i1WowbNw6nT59GeXk5ysvLcfr0aYwbNw5arRbe3t64cOEChBBPfZbsMbrbtrx79y7u378PAKisrMTXX3+NAQMG2KxNp9OpV/+//PJLjB49GoqiNHn8ssfYv39/dVsKIepsy7a2v7aEW01HmZKSgry8PBiNRvj5+eGdd95BXl4erly5AgDo06cP0tPT1b+QW7duxYEDB6DRaLB06VJEREQAAE6cOIGMjAxYLBZMnToV8+fPB1Bze9B7772H8vJy/PznP8e6devg5eXl1DH+73//wxtvvIHBgwfDw8NDHXdoaCgWLVqEW7duoXfv3ti0aRO6d+8OIQTS09Nx6tQpdO7cGRkZGRg2bBgA4NNPP8X27dsBAG+99RamTp0KALh06RLS0tJQWVmJCRMmYNmyZU79WmdrjIcOHXKrbXnlyhWkpqbCYrFACIFXXnkFCxcutFmbyWTCn/70JxQUFMDHxwcbN25EYGBgs8Yve4yzZ8+G0WiEEALBwcH44IMP4O3t3Sb315ZwqwAmImpL3P4UBBGRq2IAExFJwgAmIpKEAUxEJAkDmIhIEgYwEZEkDGBqtiFDhuDhw4eyy2jTdu3ahTt37qiv9+3bh127dskriJzKU3YBRLZYLBZoNBrZZTjU3//+d4wZMwZ+fn4AauYuofaDAUwtsnPnTuTm5sJoNCIlJUWdRvHixYtYt26deoScnJyMl156CUDN01lbt25FVVUVOnTogLS0NISFheHs2bPIyMjAr371K1y6dAnz58/HxIkT6+33+PHjyMzMhNlshoeHB1atWoXg4GCcPHkSGzZsgMViga+vL9LT0xEUFKR+9gsvvIDz589DURRs3LgRAwYMwPXr15GWlobHjx/DarUiISEBc+fORWpqKoYOHYpZs2YBwFOvU1NT4eXlpT7qGxkZiYkTJyIzMxOlpaWYM2cO5syZA6Dm8dqYmBicO3cOBoMBc+bMwaxZs7B161YYDAYkJyejY8eOWL9+PY4cOYJHjx5hyZIlsFgsWLdunTrB+vjx47F48WJoNJqn+i8tLUVYWBhWr17dZp4Aox85fsphcleDBw8WmZmZQgghrl27JkaOHClu374tysvLRVxcnNDr9UKImgnWx48fL8rLy0VRUZGYPn26qKioEEII8f3334uIiAghhBBnzpwRwcHB4ty5cw32e/36dTFmzBhx48YNIYQQJpNJVFRUiNu3b4tRo0apk47/85//VCf0PnPmjAgJCRHffvutEEKILVu2iJSUFCGEECtWrBAfffSR+vm1E30vWbJE7NmzR21/8vWSJUvEzJkzhclkEo8ePRKjR48WqampwmKxiNLSUhEWFqZORD5x4kSRmpoqhBCirKxMjB07VhQUFKg/++6779Q+Nm/eLFatWiWEEGLv3r1izpw5wmQyCZPJJGbPnq1Ozl7bf2VlpTCZTOI3v/mNOH36tL1NRi6GR8DUIomJiQCA/v37IyQkBBcuXICnpydKSkrwhz/8QX2foigoKipCfn4+iouL8cYbb6g/M5vNuH37NgAgKCgIw4cPb7DPr7/+GhMmTEDfvn0BAF5eXvDy8kJeXh6Cg4MxcOBAAMDUqVPxwQcf4MGDBwCAfv36ISQkBAAQFhaG48ePA6iZyH/16tWorq7GqFGjMHr06EaN/eWXX1bnj+jXrx8iIiLg4eGBgIAAdOvWDaWlpRgwYAAAYNq0aQBqZnR76aWX1Fob8s033yAhIUHtY8qUKTh27Bhef/11tf+OHTsCAEJCQlBcXIyxY8c2qnZyDQxgajVCCHWO1iFDhmDv3r113pOfn4/x48djzZo1dX527do1PPfcc43qp6H+bXlysh0PDw91ysfo6GiEhYUhNzcXf/vb33DgwAGsW7cOGo0GVqtV/R2TyfTU59WGH1Az5+2zry0WS7PqbOh9T75ubH/kungXBLXIgQMHAACFhYXqyhXDhw9HUVERzpw5o74vPz8fQgh1KZmrV68+9bOmGDduHE6ePInCwkIANYtV1i5LU1BQgGvXrgGomcoxJCSkwQnrAaCoqAj+/v6YMmUKFixYoM5N+/zzz6v/NhgMOHv2bJPqfFLtNJJ3797FyZMnMXLkSACAt7c3Kioq6v2dMWPGICsrC9XV1aiurkZ2djbCw8ObXQO5Hh4BU4t4eXlh5syZMBqNSE9PV6/mb9myBWvXrkVGRgaqq6sRGBiIbdu2oW/fvli7di3ef/99VFZWorq6Gi+++CJCQ0Mb3Wffvn2xYsUKvPfee+qdEqtWrcKQIUOwZs0aLF68GGazGb6+vli7dq3dzzty5Aj+/e9/o0OHDlAUBUuXLgUATJ8+HcnJyXj11VfRt2/fJtX4rN69e+P1119HWVkZ/vjHP2LIkCEAgNmzZ2Pp0qXo1KkT1q9f/9TvzJgxA8XFxeo6cOPGjcP06dObXQO5Hk5HSeRgOp0O27Ztw+DBg2WXQi6GpyCIiCThETC5rOXLl+PixYtPtWk0Ghw8eFBSRUStiwFMRCQJT0EQEUnCACYikoQBTEQkCQOYiEiS/weJsHLvNpsriAAAAABJRU5ErkJggg==\n", 
                        "text/plain": "<matplotlib.figure.Figure at 0x7f5cc1d9cd30>"
                    }, 
                    "metadata": {}
                }
            ], 
            "source": "sns.lmplot(\"beer_consumption\", \"temp_avg\", df_beer,\n          scatter_kws={\"marker\":\"x\", \"color\":\"blue\"},\n          line_kws={\"linewidth\":1, \"color\": \"orange\"})"
        }, 
        {
            "execution_count": 14, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 14, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "<seaborn.axisgrid.FacetGrid at 0x7f5cc00d7e10>"
                    }, 
                    "output_type": "execute_result"
                }, 
                {
                    "output_type": "display_data", 
                    "data": {
                        "image/png": "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\n", 
                        "text/plain": "<matplotlib.figure.Figure at 0x7f5cc00c5f28>"
                    }, 
                    "metadata": {}
                }
            ], 
            "source": "sns.lmplot(\"beer_consumption\", \"temp_max\", df_beer,\n          scatter_kws={\"marker\":\"x\", \"color\":\"blue\"},\n          line_kws={\"linewidth\":1, \"color\": \"orange\"})"
        }, 
        {
            "execution_count": 15, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 15, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "<seaborn.axisgrid.FacetGrid at 0x7f5cc007fb70>"
                    }, 
                    "output_type": "execute_result"
                }, 
                {
                    "output_type": "display_data", 
                    "data": {
                        "image/png": "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\n", 
                        "text/plain": "<matplotlib.figure.Figure at 0x7f5cc007f710>"
                    }, 
                    "metadata": {}
                }
            ], 
            "source": "sns.lmplot(\"precip\", \"beer_consumption\", df_beer,\n          scatter_kws={\"marker\":\"x\", \"color\":\"blue\"},\n          line_kws={\"linewidth\":1, \"color\": \"orange\"})"
        }, 
        {
            "source": "### Mapear dados em label para num\u00e9ricos", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 16, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "class_weekend = {True: 1, False: 0}\ndf_beer[\"weekend\"] = df_beer[\"weekend\"].map(class_weekend)"
        }, 
        {
            "execution_count": 17, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 17, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>42005.0</td>\n      <td>27.30</td>\n      <td>23.9</td>\n      <td>32.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>25461.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>42006.0</td>\n      <td>27.02</td>\n      <td>24.5</td>\n      <td>33.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>28972.0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>42007.0</td>\n      <td>24.82</td>\n      <td>22.4</td>\n      <td>29.9</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>30814.0</td>\n    </tr>\n    <tr>\n      <th>3</th>\n      <td>42008.0</td>\n      <td>23.98</td>\n      <td>21.5</td>\n      <td>28.6</td>\n      <td>1.2</td>\n      <td>1.0</td>\n      <td>29799.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>42009.0</td>\n      <td>23.82</td>\n      <td>21.0</td>\n      <td>28.3</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>28900.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "      data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n0  42005.0     27.30      23.9      32.5     0.0      0.0           25461.0\n1  42006.0     27.02      24.5      33.5     0.0      0.0           28972.0\n2  42007.0     24.82      22.4      29.9     0.0      1.0           30814.0\n3  42008.0     23.98      21.5      28.6     1.2      1.0           29799.0\n4  42009.0     23.82      21.0      28.3     0.0      0.0           28900.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.head(5)"
        }, 
        {
            "source": "### Busca por dados nulos e inv\u00e1lidos", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 18, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 18, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "data                False\ntemp_avg             True\ntemp_min             True\ntemp_max             True\nprecip              False\nweekend              True\nbeer_consumption    False\ndtype: bool"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.isnull().any()"
        }, 
        {
            "execution_count": 19, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 19, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>168</th>\n      <td>42173.0</td>\n      <td>NaN</td>\n      <td>15.8</td>\n      <td>26.2</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>24534.0</td>\n    </tr>\n    <tr>\n      <th>181</th>\n      <td>42186.0</td>\n      <td>NaN</td>\n      <td>16.2</td>\n      <td>20.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20824.0</td>\n    </tr>\n    <tr>\n      <th>309</th>\n      <td>42314.0</td>\n      <td>NaN</td>\n      <td>18.0</td>\n      <td>22.8</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20575.0</td>\n    </tr>\n    <tr>\n      <th>314</th>\n      <td>42319.0</td>\n      <td>NaN</td>\n      <td>19.8</td>\n      <td>32.7</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>29569.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n168  42173.0       NaN      15.8      26.2     0.0      0.0           24534.0\n181  42186.0       NaN      16.2      20.5     0.0      0.0           20824.0\n309  42314.0       NaN      18.0      22.8     0.0      0.0           20575.0\n314  42319.0       NaN      19.8      32.7     0.0      0.0           29569.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"temp_avg\"].isnull()]"
        }, 
        {
            "execution_count": 20, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 20, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>7</th>\n      <td>42012.0</td>\n      <td>24.90</td>\n      <td>NaN</td>\n      <td>32.8</td>\n      <td>48.6</td>\n      <td>0.0</td>\n      <td>28397.0</td>\n    </tr>\n    <tr>\n      <th>116</th>\n      <td>42121.0</td>\n      <td>19.82</td>\n      <td>NaN</td>\n      <td>24.9</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>21838.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n7    42012.0     24.90       NaN      32.8    48.6      0.0           28397.0\n116  42121.0     19.82       NaN      24.9     0.0      0.0           21838.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"temp_min\"].isnull()]"
        }, 
        {
            "execution_count": 21, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 21, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>98</th>\n      <td>42103.0</td>\n      <td>19.40</td>\n      <td>15.9</td>\n      <td>NaN</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20298.0</td>\n    </tr>\n    <tr>\n      <th>165</th>\n      <td>42170.0</td>\n      <td>16.02</td>\n      <td>13.1</td>\n      <td>NaN</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>19119.0</td>\n    </tr>\n    <tr>\n      <th>237</th>\n      <td>42242.0</td>\n      <td>18.92</td>\n      <td>14.8</td>\n      <td>NaN</td>\n      <td>0.6</td>\n      <td>0.0</td>\n      <td>23357.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n98   42103.0     19.40      15.9       NaN     0.0      0.0           20298.0\n165  42170.0     16.02      13.1       NaN     0.0      0.0           19119.0\n237  42242.0     18.92      14.8       NaN     0.6      0.0           23357.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"temp_max\"].isnull()]"
        }, 
        {
            "execution_count": 22, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 22, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>21</th>\n      <td>42026.0</td>\n      <td>21.74</td>\n      <td>19.2</td>\n      <td>0.0</td>\n      <td>31.0</td>\n      <td>NaN</td>\n      <td>25795.0</td>\n    </tr>\n    <tr>\n      <th>27</th>\n      <td>42032.0</td>\n      <td>25.68</td>\n      <td>20.1</td>\n      <td>29.9</td>\n      <td>4.9</td>\n      <td>NaN</td>\n      <td>22603.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "       data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n21  42026.0     21.74      19.2       0.0    31.0      NaN           25795.0\n27  42032.0     25.68      20.1      29.9     4.9      NaN           22603.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"weekend\"].isnull()]"
        }, 
        {
            "source": "Temperatura m\u00e9dia pode ser calculada entre a temperatura m\u00e1xima e m\u00ednima", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 23, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 23, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>168</th>\n      <td>42173.0</td>\n      <td>NaN</td>\n      <td>15.8</td>\n      <td>26.2</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>24534.0</td>\n    </tr>\n    <tr>\n      <th>181</th>\n      <td>42186.0</td>\n      <td>NaN</td>\n      <td>16.2</td>\n      <td>20.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20824.0</td>\n    </tr>\n    <tr>\n      <th>309</th>\n      <td>42314.0</td>\n      <td>NaN</td>\n      <td>18.0</td>\n      <td>22.8</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20575.0</td>\n    </tr>\n    <tr>\n      <th>314</th>\n      <td>42319.0</td>\n      <td>NaN</td>\n      <td>19.8</td>\n      <td>32.7</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>29569.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n168  42173.0       NaN      15.8      26.2     0.0      0.0           24534.0\n181  42186.0       NaN      16.2      20.5     0.0      0.0           20824.0\n309  42314.0       NaN      18.0      22.8     0.0      0.0           20575.0\n314  42319.0       NaN      19.8      32.7     0.0      0.0           29569.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer_temp_avg_null = df_beer[df_beer[\"temp_avg\"].isnull()].copy()\ndf_beer_temp_avg_null"
        }, 
        {
            "execution_count": 24, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 24, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>168</th>\n      <td>42173.0</td>\n      <td>21.00</td>\n      <td>15.8</td>\n      <td>26.2</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>24534.0</td>\n    </tr>\n    <tr>\n      <th>181</th>\n      <td>42186.0</td>\n      <td>18.35</td>\n      <td>16.2</td>\n      <td>20.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20824.0</td>\n    </tr>\n    <tr>\n      <th>309</th>\n      <td>42314.0</td>\n      <td>20.40</td>\n      <td>18.0</td>\n      <td>22.8</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20575.0</td>\n    </tr>\n    <tr>\n      <th>314</th>\n      <td>42319.0</td>\n      <td>26.25</td>\n      <td>19.8</td>\n      <td>32.7</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>29569.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n168  42173.0     21.00      15.8      26.2     0.0      0.0           24534.0\n181  42186.0     18.35      16.2      20.5     0.0      0.0           20824.0\n309  42314.0     20.40      18.0      22.8     0.0      0.0           20575.0\n314  42319.0     26.25      19.8      32.7     0.0      0.0           29569.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer_temp_avg_null[\"temp_avg\"] = (df_beer[\"temp_max\"] + df_beer[\"temp_min\"])/2\ndf_beer_temp_avg_null"
        }, 
        {
            "execution_count": 25, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "df_beer[\"temp_avg\"] = df_beer[\"temp_avg\"].replace(np.nan,(df_beer[\"temp_max\"] + df_beer[\"temp_min\"])/2)"
        }, 
        {
            "execution_count": 26, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 26, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>168</th>\n      <td>42173.0</td>\n      <td>21.00</td>\n      <td>15.8</td>\n      <td>26.2</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>24534.0</td>\n    </tr>\n    <tr>\n      <th>181</th>\n      <td>42186.0</td>\n      <td>18.35</td>\n      <td>16.2</td>\n      <td>20.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20824.0</td>\n    </tr>\n    <tr>\n      <th>309</th>\n      <td>42314.0</td>\n      <td>20.40</td>\n      <td>18.0</td>\n      <td>22.8</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20575.0</td>\n    </tr>\n    <tr>\n      <th>314</th>\n      <td>42319.0</td>\n      <td>26.25</td>\n      <td>19.8</td>\n      <td>32.7</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>29569.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n168  42173.0     21.00      15.8      26.2     0.0      0.0           24534.0\n181  42186.0     18.35      16.2      20.5     0.0      0.0           20824.0\n309  42314.0     20.40      18.0      22.8     0.0      0.0           20575.0\n314  42319.0     26.25      19.8      32.7     0.0      0.0           29569.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.loc[[168,181,309,314]]"
        }, 
        {
            "execution_count": 27, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "df_beer[\"temp_min\"] = df_beer[\"temp_min\"].replace(np.nan,df_beer[\"temp_min\"].mean())"
        }, 
        {
            "execution_count": 28, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 28, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>7</th>\n      <td>42012.0</td>\n      <td>24.90</td>\n      <td>17.39697</td>\n      <td>32.8</td>\n      <td>48.6</td>\n      <td>0.0</td>\n      <td>28397.0</td>\n    </tr>\n    <tr>\n      <th>116</th>\n      <td>42121.0</td>\n      <td>19.82</td>\n      <td>17.39697</td>\n      <td>24.9</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>21838.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n7    42012.0     24.90  17.39697      32.8    48.6      0.0           28397.0\n116  42121.0     19.82  17.39697      24.9     0.0      0.0           21838.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.loc[[7,116]]"
        }, 
        {
            "execution_count": 29, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "df_beer[\"temp_max\"] = df_beer[\"temp_max\"].replace(np.nan,df_beer[\"temp_max\"].mean())"
        }, 
        {
            "execution_count": 30, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 30, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>98</th>\n      <td>42103.0</td>\n      <td>19.40</td>\n      <td>15.9</td>\n      <td>26.480663</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>20298.0</td>\n    </tr>\n    <tr>\n      <th>165</th>\n      <td>42170.0</td>\n      <td>16.02</td>\n      <td>13.1</td>\n      <td>26.480663</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>19119.0</td>\n    </tr>\n    <tr>\n      <th>237</th>\n      <td>42242.0</td>\n      <td>18.92</td>\n      <td>14.8</td>\n      <td>26.480663</td>\n      <td>0.6</td>\n      <td>0.0</td>\n      <td>23357.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min   temp_max  precip  weekend  beer_consumption\n98   42103.0     19.40      15.9  26.480663     0.0      0.0           20298.0\n165  42170.0     16.02      13.1  26.480663     0.0      0.0           19119.0\n237  42242.0     18.92      14.8  26.480663     0.6      0.0           23357.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.loc[[98, 165, 237]]"
        }, 
        {
            "execution_count": 31, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "df_beer[\"weekend\"] = df_beer[\"weekend\"].replace(np.nan,df_beer[\"weekend\"].mode()[0])"
        }, 
        {
            "execution_count": 32, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 32, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>21</th>\n      <td>42026.0</td>\n      <td>21.74</td>\n      <td>19.2</td>\n      <td>0.0</td>\n      <td>31.0</td>\n      <td>0.0</td>\n      <td>25795.0</td>\n    </tr>\n    <tr>\n      <th>27</th>\n      <td>42032.0</td>\n      <td>25.68</td>\n      <td>20.1</td>\n      <td>29.9</td>\n      <td>4.9</td>\n      <td>0.0</td>\n      <td>22603.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "       data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n21  42026.0     21.74      19.2       0.0    31.0      0.0           25795.0\n27  42032.0     25.68      20.1      29.9     4.9      0.0           22603.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.loc[[21,27]]"
        }, 
        {
            "execution_count": 33, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 33, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "0.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[\"weekend\"].mode()[0]"
        }, 
        {
            "execution_count": 34, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 34, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "data                False\ntemp_avg            False\ntemp_min            False\ntemp_max            False\nprecip              False\nweekend             False\nbeer_consumption    False\ndtype: bool"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer.isnull().any()"
        }, 
        {
            "source": "Verifcar ocorr\u00eancia de n\u00fameros iguais a 0, que seria inv\u00e1lido", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 35, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 35, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "data                False\ntemp_avg             True\ntemp_min             True\ntemp_max             True\nprecip               True\nweekend              True\nbeer_consumption    False\ndtype: bool"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "(df_beer == 0).any()"
        }, 
        {
            "execution_count": 36, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 36, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>323</th>\n      <td>42328.0</td>\n      <td>0.0</td>\n      <td>19.6</td>\n      <td>27.0</td>\n      <td>6.8</td>\n      <td>0.0</td>\n      <td>20332.0</td>\n    </tr>\n    <tr>\n      <th>339</th>\n      <td>42344.0</td>\n      <td>0.0</td>\n      <td>20.6</td>\n      <td>28.0</td>\n      <td>0.1</td>\n      <td>1.0</td>\n      <td>32780.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n323  42328.0       0.0      19.6      27.0     6.8      0.0           20332.0\n339  42344.0       0.0      20.6      28.0     0.1      1.0           32780.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"temp_avg\"] == 0]"
        }, 
        {
            "execution_count": 37, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 37, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>13</th>\n      <td>42018.0</td>\n      <td>25.96</td>\n      <td>0.0</td>\n      <td>34.0</td>\n      <td>1.6</td>\n      <td>0.0</td>\n      <td>31825.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "       data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n13  42018.0     25.96       0.0      34.0     1.6      0.0           31825.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"temp_min\"] == 0]"
        }, 
        {
            "execution_count": 38, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 38, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>21</th>\n      <td>42026.0</td>\n      <td>21.74</td>\n      <td>19.2</td>\n      <td>0.0</td>\n      <td>31.0</td>\n      <td>0.0</td>\n      <td>25795.0</td>\n    </tr>\n    <tr>\n      <th>113</th>\n      <td>42118.0</td>\n      <td>22.52</td>\n      <td>19.3</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>24388.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "        data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n21   42026.0     21.74      19.2       0.0    31.0      0.0           25795.0\n113  42118.0     22.52      19.3       0.0     0.0      0.0           24388.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"temp_max\"] == 0]"
        }, 
        {
            "execution_count": 39, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 39, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>42005.0</td>\n      <td>27.30</td>\n      <td>23.9</td>\n      <td>32.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>25461.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>42006.0</td>\n      <td>27.02</td>\n      <td>24.5</td>\n      <td>33.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>28972.0</td>\n    </tr>\n    <tr>\n      <th>2</th>\n      <td>42007.0</td>\n      <td>24.82</td>\n      <td>22.4</td>\n      <td>29.9</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>30814.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>42009.0</td>\n      <td>23.82</td>\n      <td>21.0</td>\n      <td>28.3</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>28900.0</td>\n    </tr>\n    <tr>\n      <th>6</th>\n      <td>42011.0</td>\n      <td>24.00</td>\n      <td>19.5</td>\n      <td>33.7</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>29732.0</td>\n    </tr>\n    <tr>\n      <th>9</th>\n      <td>42014.0</td>\n      <td>26.76</td>\n      <td>22.1</td>\n      <td>34.2</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>37937.0</td>\n    </tr>\n    <tr>\n      <th>11</th>\n      <td>42016.0</td>\n      <td>25.96</td>\n      <td>21.4</td>\n      <td>35.4</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>25743.0</td>\n    </tr>\n    <tr>\n      <th>16</th>\n      <td>42021.0</td>\n      <td>28.86</td>\n      <td>22.0</td>\n      <td>35.8</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>37690.0</td>\n    </tr>\n    <tr>\n      <th>17</th>\n      <td>42022.0</td>\n      <td>28.26</td>\n      <td>23.4</td>\n      <td>35.6</td>\n      <td>0.0</td>\n      <td>1.0</td>\n      <td>30524.0</td>\n    </tr>\n    <tr>\n      <th>20</th>\n      <td>42025.0</td>\n      <td>25.32</td>\n      <td>22.7</td>\n      <td>30.9</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>29130.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "       data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n0   42005.0     27.30      23.9      32.5     0.0      0.0           25461.0\n1   42006.0     27.02      24.5      33.5     0.0      0.0           28972.0\n2   42007.0     24.82      22.4      29.9     0.0      1.0           30814.0\n4   42009.0     23.82      21.0      28.3     0.0      0.0           28900.0\n6   42011.0     24.00      19.5      33.7     0.0      0.0           29732.0\n9   42014.0     26.76      22.1      34.2     0.0      1.0           37937.0\n11  42016.0     25.96      21.4      35.4     0.0      0.0           25743.0\n16  42021.0     28.86      22.0      35.8     0.0      1.0           37690.0\n17  42022.0     28.26      23.4      35.6     0.0      1.0           30524.0\n20  42025.0     25.32      22.7      30.9     0.0      0.0           29130.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"precip\"] == 0].head(10)"
        }, 
        {
            "execution_count": 40, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 40, 
                    "metadata": {}, 
                    "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>data</th>\n      <th>temp_avg</th>\n      <th>temp_min</th>\n      <th>temp_max</th>\n      <th>precip</th>\n      <th>weekend</th>\n      <th>beer_consumption</th>\n    </tr>\n  </thead>\n  <tbody>\n    <tr>\n      <th>0</th>\n      <td>42005.0</td>\n      <td>27.30</td>\n      <td>23.90000</td>\n      <td>32.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>25461.0</td>\n    </tr>\n    <tr>\n      <th>1</th>\n      <td>42006.0</td>\n      <td>27.02</td>\n      <td>24.50000</td>\n      <td>33.5</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>28972.0</td>\n    </tr>\n    <tr>\n      <th>4</th>\n      <td>42009.0</td>\n      <td>23.82</td>\n      <td>21.00000</td>\n      <td>28.3</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>28900.0</td>\n    </tr>\n    <tr>\n      <th>5</th>\n      <td>42010.0</td>\n      <td>23.78</td>\n      <td>20.10000</td>\n      <td>30.5</td>\n      <td>12.2</td>\n      <td>0.0</td>\n      <td>28218.0</td>\n    </tr>\n    <tr>\n      <th>6</th>\n      <td>42011.0</td>\n      <td>24.00</td>\n      <td>19.50000</td>\n      <td>33.7</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>29732.0</td>\n    </tr>\n    <tr>\n      <th>7</th>\n      <td>42012.0</td>\n      <td>24.90</td>\n      <td>17.39697</td>\n      <td>32.8</td>\n      <td>48.6</td>\n      <td>0.0</td>\n      <td>28397.0</td>\n    </tr>\n    <tr>\n      <th>8</th>\n      <td>42013.0</td>\n      <td>28.20</td>\n      <td>21.90000</td>\n      <td>34.0</td>\n      <td>4.4</td>\n      <td>0.0</td>\n      <td>24886.0</td>\n    </tr>\n    <tr>\n      <th>11</th>\n      <td>42016.0</td>\n      <td>25.96</td>\n      <td>21.40000</td>\n      <td>35.4</td>\n      <td>0.0</td>\n      <td>0.0</td>\n      <td>25743.0</td>\n    </tr>\n    <tr>\n      <th>12</th>\n      <td>42017.0</td>\n      <td>25.52</td>\n      <td>21.20000</td>\n      <td>34.8</td>\n      <td>0.8</td>\n      <td>0.0</td>\n      <td>26990.0</td>\n    </tr>\n    <tr>\n      <th>13</th>\n      <td>42018.0</td>\n      <td>25.96</td>\n      <td>0.00000</td>\n      <td>34.0</td>\n      <td>1.6</td>\n      <td>0.0</td>\n      <td>31825.0</td>\n    </tr>\n  </tbody>\n</table>\n</div>", 
                        "text/plain": "       data  temp_avg  temp_min  temp_max  precip  weekend  beer_consumption\n0   42005.0     27.30  23.90000      32.5     0.0      0.0           25461.0\n1   42006.0     27.02  24.50000      33.5     0.0      0.0           28972.0\n4   42009.0     23.82  21.00000      28.3     0.0      0.0           28900.0\n5   42010.0     23.78  20.10000      30.5    12.2      0.0           28218.0\n6   42011.0     24.00  19.50000      33.7     0.0      0.0           29732.0\n7   42012.0     24.90  17.39697      32.8    48.6      0.0           28397.0\n8   42013.0     28.20  21.90000      34.0     4.4      0.0           24886.0\n11  42016.0     25.96  21.40000      35.4     0.0      0.0           25743.0\n12  42017.0     25.52  21.20000      34.8     0.8      0.0           26990.0\n13  42018.0     25.96   0.00000      34.0     1.6      0.0           31825.0"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "df_beer[df_beer[\"weekend\"] == 0].head(10)"
        }, 
        {
            "source": "Precipita\u00e7\u00e3o e Weekend podem ser igual a 0.", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "Utilizar mesma estrat\u00e9gia anterior para os demais atributos.", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 41, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "df_beer[\"temp_avg\"] = df_beer[\"temp_avg\"].replace(0,(df_beer[\"temp_max\"] + df_beer[\"temp_min\"])/2)\ndf_beer[\"temp_min\"] = df_beer[\"temp_min\"].replace(0,df_beer[\"temp_min\"].mean())\ndf_beer[\"temp_max\"] = df_beer[\"temp_max\"].replace(0,df_beer[\"temp_max\"].mean())"
        }, 
        {
            "execution_count": 42, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 42, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "data                False\ntemp_avg            False\ntemp_min            False\ntemp_max            False\nprecip               True\nweekend              True\nbeer_consumption    False\ndtype: bool"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "(df_beer == 0).any()"
        }, 
        {
            "source": "### Seperando dados de treinamento e teste", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 43, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [], 
            "source": "from sklearn.model_selection import train_test_split\n\nfeature_col_names = ['temp_max', 'precip', 'weekend']\npredicted_class_names = ['beer_consumption']\n\nX = df_beer[feature_col_names].values\ny = df_beer[predicted_class_names].values\nsplit_test_size = 0.30\n\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=split_test_size, random_state=42)"
        }, 
        {
            "execution_count": 44, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "output_type": "stream", 
                    "name": "stdout", 
                    "text": "69.86% in training set\n30.14% in test set\n"
                }
            ], 
            "source": "print(\"{0:0.2f}% in training set\".format((len(X_train)/len(df_beer.index)) * 100))\nprint(\"{0:0.2f}% in test set\".format((len(X_test)/len(df_beer.index)) * 100))"
        }, 
        {
            "execution_count": 45, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 45, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "LinearRegression(copy_X=True, fit_intercept=True, n_jobs=1, normalize=False)"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "from sklearn import linear_model\n\nlr_model = linear_model.LinearRegression()\nlr_model.fit(X_train, y_train.ravel())"
        }, 
        {
            "source": "Testando valores par verificar a performance", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 46, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "output_type": "stream", 
                    "name": "stdout", 
                    "text": "R2 score: 0.74\nR2 score: 0.74\n"
                }
            ], 
            "source": "from sklearn.metrics import mean_squared_error, r2_score\n\ny_pred = lr_model.predict(X_test)\n\nprint('R2 score: %.2f' % r2_score(y_test, y_pred))\nprint('R2 score: %.2f' % lr_model.score(X_test, y_test))"
        }, 
        {
            "source": "R2 score est\u00e1 na faixa esperada de >= 70%", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "Simula\u00e7\u00e3o de um caso onde se espera maior consumo, clima quente, 35 graus C, sem chuva e no final de semana.", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 47, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 47, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "array([ 35092.25246658])"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "predict_value = [[35, 0, 1]]\nlr_model.predict(predict_value)"
        }, 
        {
            "source": "Simula\u00e7\u00e3o de um caso onde se espera um consumo menor que o de cima, clima quente, 35 graus C, sem chuva mas em dia de semana.", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 48, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 48, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "array([ 29753.3647769])"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "predict_value = [[35, 0, 0]]\nlr_model.predict(predict_value)"
        }, 
        {
            "source": "Simula\u00e7\u00e3o de um caso onde se espera um consumo menor que o de cima, clima quente, 35 graus C, com chuva e em dia de semana.", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 49, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 49, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "array([ 28689.99450697])"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "predict_value = [[35, 20, 0]]\nlr_model.predict(predict_value)"
        }, 
        {
            "source": "Pior cen\u00e1rio, baixo consumo, frio (10 graus C), chovendo e em dia de semana.", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "execution_count": 50, 
            "cell_type": "code", 
            "metadata": {}, 
            "outputs": [
                {
                    "execution_count": 50, 
                    "metadata": {}, 
                    "data": {
                        "text/plain": "array([ 12008.88877202])"
                    }, 
                    "output_type": "execute_result"
                }
            ], 
            "source": "predict_value = [[10, 20, 0]]\nlr_model.predict(predict_value)"
        }, 
        {
            "source": "### Salvando modelo", 
            "cell_type": "markdown", 
            "metadata": {}
        }, 
        {
            "source": "from sklearn.externals import joblib\njoblib.dump(lr_model, 'models/lr_model.pkl') ", 
            "cell_type": "raw", 
            "metadata": {}
        }, 
        {
            "source": "lr_model_loaded = joblib.load('models/lr_model.pkl') ", 
            "cell_type": "raw", 
            "metadata": {}
        }, 
        {
            "source": "predict_value = [[10, 20, 0]]\nlr_model_loaded.predict(predict_value)", 
            "cell_type": "raw", 
            "metadata": {}
        }, 
        {
            "source": "Por Claudio Pinheiro \nhttps://www.linkedin.com/in/claudio-pinheiro-cloud/", 
            "cell_type": "markdown", 
            "metadata": {}
        }
    ], 
    "metadata": {
        "kernelspec": {
            "display_name": "Python 3.5", 
            "name": "python3", 
            "language": "python"
        }, 
        "language_info": {
            "mimetype": "text/x-python", 
            "nbconvert_exporter": "python", 
            "version": "3.5.5", 
            "name": "python", 
            "file_extension": ".py", 
            "pygments_lexer": "ipython3", 
            "codemirror_mode": {
                "version": 3, 
                "name": "ipython"
            }
        }
    }, 
    "nbformat": 4
}