agrupamento.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "CEtG6c4koygW"
   },
   "source": [
    "# Projeto #1 - Mineração de agrupamentos\n",
    "\n",
    "<p style='text-align: justify;'><font size=3.>Análise de acidentes nas rodovias federais brasileira entre 2007 e 2021.</font></p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "IGJiaCUD1ncA"
   },
   "source": [
    "### Bibliotecas básicas e outros imports"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:14:59.838244Z",
     "start_time": "2022-02-14T21:14:58.306502Z"
    },
    "id": "4bG6r2Bwoyge"
   },
   "outputs": [],
   "source": [
    "import warnings\n",
    "import matplotlib\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "import matplotlib.pyplot as plt\n",
    "from sklearn.cluster import KMeans\n",
    "from sklearn.metrics import silhouette_score\n",
    "from sklearn.preprocessing import Normalizer\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from pyclustertend import hopkins\n",
    "from sklearn.preprocessing import StandardScaler\n",
    "\n",
    "%run modules/text.py\n",
    "\n",
    "%matplotlib inline\n",
    "plt.rcParams['font.family'] = 'Arial, Helvetica, sans-serif'\n",
    "\n",
    "warnings.filterwarnings('ignore')\n",
    "pd.set_option('display.max_rows', 1000)\n",
    "pd.set_option('display.max_columns', 200)\n",
    "np.set_printoptions(suppress=True, precision=3)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "id": "rjtF8niV1ncF"
   },
   "source": [
    "### Carregamento dos dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:15:05.946553Z",
     "start_time": "2022-02-14T21:14:59.840755Z"
    },
    "colab": {
     "base_uri": "https://localhost:8080/",
     "height": 134
    },
    "id": "H8Xxqiwcoygh",
    "outputId": "948da79f-e337-4029-b318-a0ef40832d18"
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Linhas: 1707393 | Coluna:  27\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>id</th>\n",
       "      <th>dia_semana</th>\n",
       "      <th>uf</th>\n",
       "      <th>br</th>\n",
       "      <th>municipio</th>\n",
       "      <th>causa_acidente</th>\n",
       "      <th>tipo_acidente</th>\n",
       "      <th>classificacao_acidente</th>\n",
       "      <th>fase_dia</th>\n",
       "      <th>sentido_via</th>\n",
       "      <th>condicao_metereologica</th>\n",
       "      <th>tipo_pista</th>\n",
       "      <th>tracado_via</th>\n",
       "      <th>uso_solo</th>\n",
       "      <th>tipo_veiculo</th>\n",
       "      <th>tipo_envolvido</th>\n",
       "      <th>estado_fisico</th>\n",
       "      <th>idade</th>\n",
       "      <th>sexo</th>\n",
       "      <th>data</th>\n",
       "      <th>mes</th>\n",
       "      <th>ano</th>\n",
       "      <th>tipo_acidente_agrupado</th>\n",
       "      <th>fase_dia_agrupado</th>\n",
       "      <th>dia_semana_agrupado</th>\n",
       "      <th>condicao_metereologica_agrupado</th>\n",
       "      <th>porte_veiculo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>83123216.0</td>\n",
       "      <td>quarta</td>\n",
       "      <td>SC</td>\n",
       "      <td>101</td>\n",
       "      <td>Balneario Camboriu</td>\n",
       "      <td>nao guardar distancia de seguranca</td>\n",
       "      <td>colisao traseira</td>\n",
       "      <td>sem vitimas</td>\n",
       "      <td>plena noite</td>\n",
       "      <td>decrescente</td>\n",
       "      <td>chuva</td>\n",
       "      <td>simples</td>\n",
       "      <td>reta</td>\n",
       "      <td>urbano</td>\n",
       "      <td>automovel</td>\n",
       "      <td>condutor</td>\n",
       "      <td>ileso</td>\n",
       "      <td>29</td>\n",
       "      <td>masculino</td>\n",
       "      <td>2014-01-01</td>\n",
       "      <td>1</td>\n",
       "      <td>2014</td>\n",
       "      <td>colisão</td>\n",
       "      <td>noite</td>\n",
       "      <td>dia útil</td>\n",
       "      <td>tempo ruim</td>\n",
       "      <td>pequeno porte</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "           id dia_semana  uf   br           municipio  \\\n",
       "0  83123216.0     quarta  SC  101  Balneario Camboriu   \n",
       "\n",
       "                       causa_acidente     tipo_acidente  \\\n",
       "0  nao guardar distancia de seguranca  colisao traseira   \n",
       "\n",
       "  classificacao_acidente     fase_dia  sentido_via condicao_metereologica  \\\n",
       "0            sem vitimas  plena noite  decrescente                  chuva   \n",
       "\n",
       "  tipo_pista tracado_via uso_solo tipo_veiculo tipo_envolvido estado_fisico  \\\n",
       "0    simples        reta   urbano    automovel       condutor         ileso   \n",
       "\n",
       "   idade       sexo        data  mes   ano tipo_acidente_agrupado  \\\n",
       "0     29  masculino  2014-01-01    1  2014                colisão   \n",
       "\n",
       "  fase_dia_agrupado dia_semana_agrupado condicao_metereologica_agrupado  \\\n",
       "0             noite            dia útil                      tempo ruim   \n",
       "\n",
       "   porte_veiculo  \n",
       "0  pequeno porte  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df = pd.read_csv('dataset/cluster_acidentes_rodovias_brasileiras_2007_a_2021.csv', low_memory=False, encoding='utf-8')\n",
    "print(f'Linhas: {df.shape[0]} | Coluna:  {df.shape[1]}')\n",
    "df.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Categorização da variável idade\n",
    "\n",
    "> <p style='text-align: justify;'><font size=2.85>Categorização realizada de acordo com o estatuto da criança e adolescente (ECA).</font></p>\n",
    "\n",
    "| faixa | categoria |\n",
    "| --- | --- |\n",
    "| 0-11 | criança |\n",
    "| 12-17  | adolescente |\n",
    "| 18-29 | jovem |\n",
    "| 30-59 | adulto |\n",
    "| acima de 60 | idoso |"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:15:06.002904Z",
     "start_time": "2022-02-14T21:15:05.948534Z"
    }
   },
   "outputs": [],
   "source": [
    "labels = ['criança', 'adolescente', 'jovem', 'adulto', 'idoso']\n",
    "classes = [0, 11, 17, 29, 59, 99]\n",
    "df['faixa_etaria'] = pd.cut(x=df['idade'], bins=classes, labels=labels, include_lowest=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Seleção de features"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:15:06.116097Z",
     "start_time": "2022-02-14T21:15:06.005407Z"
    },
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>fase_dia_agrupado</th>\n",
       "      <th>tipo_acidente</th>\n",
       "      <th>causa_acidente</th>\n",
       "      <th>condicao_metereologica</th>\n",
       "      <th>tracado_via</th>\n",
       "      <th>classificacao_acidente</th>\n",
       "      <th>dia_semana</th>\n",
       "      <th>tipo_pista</th>\n",
       "      <th>porte_veiculo</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>noite</td>\n",
       "      <td>colisao traseira</td>\n",
       "      <td>nao guardar distancia de seguranca</td>\n",
       "      <td>chuva</td>\n",
       "      <td>reta</td>\n",
       "      <td>sem vitimas</td>\n",
       "      <td>quarta</td>\n",
       "      <td>simples</td>\n",
       "      <td>pequeno porte</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  fase_dia_agrupado     tipo_acidente                      causa_acidente  \\\n",
       "0             noite  colisao traseira  nao guardar distancia de seguranca   \n",
       "\n",
       "  condicao_metereologica tracado_via classificacao_acidente dia_semana  \\\n",
       "0                  chuva        reta            sem vitimas     quarta   \n",
       "\n",
       "  tipo_pista  porte_veiculo  \n",
       "0    simples  pequeno porte  "
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "colunas = [\n",
    "    'fase_dia_agrupado',\n",
    "    'tipo_acidente',\n",
    "    'causa_acidente',\n",
    "    'condicao_metereologica',\n",
    "    'tracado_via',\n",
    "    'classificacao_acidente',\n",
    "    'dia_semana',\n",
    "    'tipo_pista',\n",
    "    'porte_veiculo',\n",
    "]\n",
    "\n",
    "dados = df[colunas]\n",
    "dados.head(1)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "lang": "en"
   },
   "source": [
    "### Conversão para valores numéricos"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:15:08.402702Z",
     "start_time": "2022-02-14T21:15:06.117747Z"
    }
   },
   "outputs": [],
   "source": [
    "le = LabelEncoder()\n",
    "dados_copy = pd.DataFrame()\n",
    "for col in dados.columns:\n",
    "    dados_copy[col+'_encoder'] = le.fit_transform(dados[col].astype(str))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Normalização dos dados"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:15:08.670219Z",
     "start_time": "2022-02-14T21:15:08.405655Z"
    }
   },
   "outputs": [],
   "source": [
    "norm = Normalizer(norm='l2')\n",
    "scaled = norm.fit_transform(dados_copy)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Tendência de agrupamento - Estatística de Hopkins\n",
    "\n",
    "<p style='text-align: justify;'><font size=2.85>H⁰: conjunto de dados é uniformemente distribuído e não contém clusters significativos.</font></p>\n",
    "\n",
    "<p style='text-align: justify;'><font size=2.85>H¹: conjunto de dados não está uniformemente distribuído e contém clusters significativos.</font></p>\n",
    "\n",
    "> <p style='text-align: justify;'><font size=2.85>Se H < 0.5, então não rejeita H⁰, pois é improvável que o conjunto tenha clusters estatisticamente significativos. Se o valor da estatística de Hopkins for próximo a 1, então podemos rejeitar a H⁰ e concluir que o conjunto de dados é significativamente um dado agrupável.</font></p>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:16:25.118448Z",
     "start_time": "2022-02-14T21:15:08.672441Z"
    },
    "scrolled": false
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Estatística de hopkins: 0.31 +/- 0.01\n"
     ]
    }
   ],
   "source": [
    "scaled = pd.DataFrame(scaled)\n",
    "scaled.columns = dados_copy.columns\n",
    "\n",
    "stats = []\n",
    "for i in range(0, 1025):\n",
    "    sample = dados_copy.sample(n=500, replace=False)\n",
    "    stats.append(hopkins(sample, sample.shape[0]))\n",
    "print(f'Estatística de hopkins: {np.mean(stats).round(2)} +/- {np.std(stats).round(2)}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<p style='text-align: justify;'><font size=2.85>Pelo que podemos perceber pela estatística de hopkins, é improvável que esse conjunto de dados seja agrupável, ou seja, tenha clusters significativos. Nesse caso, a clusterização não é a técnica ideal para lidar com esse problema, mas para fins didáticos seguiremos com as análises.</font></p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Mineração de agrupamento"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Método Elbow (cotovelo)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:19:42.367178Z",
     "start_time": "2022-02-14T21:16:25.120058Z"
    }
   },
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "inertias = []\n",
    "ks = range(1, 11)\n",
    "\n",
    "for k in ks:\n",
    "    model = KMeans(n_clusters=k, init='k-means++', random_state=42)\n",
    "    model.fit(scaled)\n",
    "    inertias.append(model.inertia_)\n",
    "    \n",
    "\n",
    "suptitle = 'Método Elbow'\n",
    "title = 'Seleção do número de clusters para o algoritmo K-means'\n",
    "\n",
    "plt.figure(figsize=(12, 6))\n",
    "plt.gca().spines['top'].set_visible(False)\n",
    "plt.gca().spines['right'].set_visible(False)\n",
    "plt.plot(ks, inertias, '-o', color='#191970')\n",
    "plt.suptitle(suptitle, fontsize=18, y=1.02, horizontalalignment='left', x=0.12, color='black', weight='bold')\n",
    "plt.title(title, fontsize=14, color='#646464', y=1.05, loc='left', x=-0.01)\n",
    "plt.xlabel(None)\n",
    "plt.ylabel('soma das distâncias quadradas', fontsize=14, horizontalalignment='left', y=0.32)\n",
    "plt.xticks(ks)\n",
    "plt.grid(False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Coeficiente de Silhouette\n",
    "\n",
    "> <p style='text-align: justify;'><font size=2.85>É uma medida de quão semelhante um ponto de dados é dentro do cluster (coesão) em comparação com outros clusters (separação). O melhor valor é 1 e o pior valor é -1. Valores próximos a 0 indicam clusters sobrepostos. Valores negativos geralmente indicam que uma amostra foi atribuída ao cluster errado, pois um cluster diferente é mais semelhante.</font></p>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:31:00.991093Z",
     "start_time": "2022-02-14T21:19:42.369223Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Para n_clusters = 2 a média do silhouette_score é : 0.5207287977268694\n",
      "Para n_clusters = 3 a média do silhouette_score é : 0.5351882784915003\n",
      "Para n_clusters = 4 a média do silhouette_score é : 0.3140714673105126\n",
      "Para n_clusters = 5 a média do silhouette_score é : 0.27341211368336504\n"
     ]
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x432 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "randomized = dados_copy.iloc[np.random.permutation(dados_copy.index)].reset_index(drop=True)\n",
    "sample = randomized.sample(n=100000)\n",
    "X = norm.fit_transform(sample)\n",
    "\n",
    "range_n_clusters = range(2, 6)\n",
    "silhouette_avg = []\n",
    "\n",
    "for k in range_n_clusters:\n",
    "    model = KMeans(n_clusters=k, init='k-means++', random_state=0)\n",
    "    cluster_labels = model.fit_predict(X)\n",
    "    score = silhouette_score(X, cluster_labels)\n",
    "    silhouette_avg.append(score)\n",
    "    print('Para n_clusters =', k, 'a média do silhouette_score é :', score)\n",
    "\n",
    "\n",
    "suptitle = 'Coeficiente de Silhouette'\n",
    "title = 'Seleção do número de clusters para o algoritmo K-means'\n",
    "\n",
    "plt.figure(figsize=(12, 6))\n",
    "plt.suptitle(suptitle, fontsize=18, y=1.02, horizontalalignment='left', x=0.12, color='black', weight='bold')\n",
    "plt.title(title, fontsize=14, color='#646464', y=1.05, loc='left', x=-0.01)\n",
    "plt.plot(range_n_clusters, silhouette_avg, 'bx-', color='#191970')\n",
    "plt.gca().spines['top'].set_visible(False)\n",
    "plt.gca().spines['right'].set_visible(False)\n",
    "plt.xlabel(None)\n",
    "plt.ylabel('Coeficiente Silhouette', fontsize=14, horizontalalignment='left', y=0.51) \n",
    "plt.grid(False)\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "<p style='text-align: justify;'><font size=2.85>A maior média do coeficiente de silhouette é 0.53 para três grupos. Esse valor é um indicativo da existência de clusters sobrepostos. Um valor baixo já era esperado, uma vez que na análise anterior percebemos que os dados não são agrupáveis.</font></p>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "lang": "fr"
   },
   "source": [
    "Agrupamento"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:31:18.904726Z",
     "start_time": "2022-02-14T21:31:00.992607Z"
    }
   },
   "outputs": [],
   "source": [
    "kmeans= KMeans(n_clusters=3, init='k-means++', random_state=0).fit(scaled)\n",
    "dados['cluster'] = kmeans.labels_"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Análise do agrupamento"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:31:18.925287Z",
     "start_time": "2022-02-14T21:31:18.906260Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "0    1475185\n",
      "1     194348\n",
      "2      37860\n",
      "Name: cluster, dtype: int64\n"
     ]
    }
   ],
   "source": [
    "print(dados['cluster'].value_counts())"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cluster 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:31:19.985273Z",
     "start_time": "2022-02-14T21:31:18.926501Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Coluna fase_dia_agrupado: dia\n",
      "Coluna tipo_acidente: colisao traseira\n",
      "Coluna causa_acidente: falta de atencao\n",
      "Coluna condicao_metereologica: ceu claro\n",
      "Coluna tracado_via: reta\n",
      "Coluna classificacao_acidente: sem vitimas\n",
      "Coluna dia_semana: sexta\n",
      "Coluna tipo_pista: simples\n",
      "Coluna porte_veiculo: pequeno porte\n",
      "Coluna cluster: 0\n"
     ]
    }
   ],
   "source": [
    "cluster_um = dados.query('cluster==0')\n",
    "for col in cluster_um:\n",
    "    print(f'Coluna {col}: {cluster_um[col].value_counts().index[0]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cluster 2"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:31:20.166828Z",
     "start_time": "2022-02-14T21:31:19.986614Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Coluna fase_dia_agrupado: dia\n",
      "Coluna tipo_acidente: saida de pista\n",
      "Coluna causa_acidente: falta de atencao\n",
      "Coluna condicao_metereologica: ceu claro\n",
      "Coluna tracado_via: reta\n",
      "Coluna classificacao_acidente: com vitimas feridas\n",
      "Coluna dia_semana: domingo\n",
      "Coluna tipo_pista: simples\n",
      "Coluna porte_veiculo: pequeno porte\n",
      "Coluna cluster: 1\n"
     ]
    }
   ],
   "source": [
    "cluster_dois = dados.query('cluster==1')\n",
    "for col in cluster_dois:\n",
    "    print(f'Coluna {col}: {cluster_dois[col].value_counts().index[0]}')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Cluster 3"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {
    "ExecuteTime": {
     "end_time": "2022-02-14T21:31:20.222161Z",
     "start_time": "2022-02-14T21:31:20.168071Z"
    }
   },
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Coluna fase_dia_agrupado: noite\n",
      "Coluna tipo_acidente: atropelamento de animal\n",
      "Coluna causa_acidente: animais na pista\n",
      "Coluna condicao_metereologica: ceu claro\n",
      "Coluna tracado_via: reta\n",
      "Coluna classificacao_acidente: sem vitimas\n",
      "Coluna dia_semana: sexta\n",
      "Coluna tipo_pista: simples\n",
      "Coluna porte_veiculo: pequeno porte\n",
      "Coluna cluster: 2\n"
     ]
    }
   ],
   "source": [
    "cluster_tres = dados.query('cluster==2')\n",
    "for col in cluster_tres:\n",
    "    print(f'Coluna {col}: {cluster_tres[col].value_counts().index[0]}')"
   ]
  }
 ],
 "metadata": {
  "colab": {
   "collapsed_sections": [],
   "name": "regra_associação.ipynb",
   "provenance": []
  },
  "file_extension": ".py",
  "hide_input": false,
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.9"
  },
  "mimetype": "text/x-python",
  "name": "python",
  "nbTranslate": {
   "displayLangs": [
    "*"
   ],
   "hotkey": "alt-t",
   "langInMainMenu": true,
   "sourceLang": "en",
   "targetLang": "fr",
   "useGoogleTranslate": true
  },
  "npconvert_exporter": "python",
  "pygments_lexer": "ipython3",
  "toc": {
   "base_numbering": 1,
   "nav_menu": {},
   "number_sections": true,
   "sideBar": true,
   "skip_h1_title": false,
   "title_cell": "Table of Contents",
   "title_sidebar": "Contents",
   "toc_cell": false,
   "toc_position": {},
   "toc_section_display": true,
   "toc_window_display": false
  },
  "varInspector": {
   "cols": {
    "lenName": 16,
    "lenType": 16,
    "lenVar": 40
   },
   "kernels_config": {
    "python": {
     "delete_cmd_postfix": "",
     "delete_cmd_prefix": "del ",
     "library": "var_list.py",
     "varRefreshCmd": "print(var_dic_list())"
    },
    "r": {
     "delete_cmd_postfix": ") ",
     "delete_cmd_prefix": "rm(",
     "library": "var_list.r",
     "varRefreshCmd": "cat(var_dic_list()) "
    }
   },
   "types_to_exclude": [
    "module",
    "function",
    "builtin_function_or_method",
    "instance",
    "_Feature"
   ],
   "window_display": false
  },
  "version": 3
 },
 "nbformat": 4,
 "nbformat_minor": 1
}