Model-Exploration-OpenData.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 62,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Data Exploration"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Create tables"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 63,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>playful,serious</th>\n",
       "      <th>shy,bold</th>\n",
       "      <th>cheery,sorrowful</th>\n",
       "      <th>masculine,feminine</th>\n",
       "      <th>charming,awkward</th>\n",
       "      <th>lewd,tasteful</th>\n",
       "      <th>intellectual,physical</th>\n",
       "      <th>strict,lenient</th>\n",
       "      <th>refined,rugged</th>\n",
       "      <th>trusting,suspicious</th>\n",
       "      <th>...</th>\n",
       "      <th>picky,always down</th>\n",
       "      <th>proactive,reactive</th>\n",
       "      <th>prudish,flirtatious</th>\n",
       "      <th>innocent,jaded</th>\n",
       "      <th>touchy-feely,distant</th>\n",
       "      <th>muddy,washed</th>\n",
       "      <th>quirky,predictable</th>\n",
       "      <th>never cries,often crying</th>\n",
       "      <th>race.label</th>\n",
       "      <th>gender.label</th>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>id</th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "      <th></th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>WE/1</th>\n",
       "      <td>36.4</td>\n",
       "      <td>98.8</td>\n",
       "      <td>72.3</td>\n",
       "      <td>47.3</td>\n",
       "      <td>24.2</td>\n",
       "      <td>29.3</td>\n",
       "      <td>75.4</td>\n",
       "      <td>64.7</td>\n",
       "      <td>82.1</td>\n",
       "      <td>57.2</td>\n",
       "      <td>...</td>\n",
       "      <td>60.4</td>\n",
       "      <td>74.0</td>\n",
       "      <td>83.9</td>\n",
       "      <td>78.7</td>\n",
       "      <td>70.9</td>\n",
       "      <td>33.2</td>\n",
       "      <td>31.3</td>\n",
       "      <td>31.8</td>\n",
       "      <td>white</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WE/2</th>\n",
       "      <td>57.4</td>\n",
       "      <td>89.2</td>\n",
       "      <td>64.7</td>\n",
       "      <td>13.8</td>\n",
       "      <td>19.2</td>\n",
       "      <td>45.0</td>\n",
       "      <td>67.5</td>\n",
       "      <td>38.9</td>\n",
       "      <td>80.4</td>\n",
       "      <td>54.2</td>\n",
       "      <td>...</td>\n",
       "      <td>73.9</td>\n",
       "      <td>65.0</td>\n",
       "      <td>93.3</td>\n",
       "      <td>81.0</td>\n",
       "      <td>66.0</td>\n",
       "      <td>28.1</td>\n",
       "      <td>41.7</td>\n",
       "      <td>27.5</td>\n",
       "      <td>white</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WE/3</th>\n",
       "      <td>26.8</td>\n",
       "      <td>57.4</td>\n",
       "      <td>9.1</td>\n",
       "      <td>92.1</td>\n",
       "      <td>27.1</td>\n",
       "      <td>66.8</td>\n",
       "      <td>44.0</td>\n",
       "      <td>74.3</td>\n",
       "      <td>34.0</td>\n",
       "      <td>12.3</td>\n",
       "      <td>...</td>\n",
       "      <td>71.4</td>\n",
       "      <td>44.9</td>\n",
       "      <td>76.8</td>\n",
       "      <td>31.4</td>\n",
       "      <td>13.3</td>\n",
       "      <td>84.4</td>\n",
       "      <td>12.7</td>\n",
       "      <td>72.8</td>\n",
       "      <td>white</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WE/4</th>\n",
       "      <td>56.9</td>\n",
       "      <td>74.4</td>\n",
       "      <td>28.2</td>\n",
       "      <td>58.4</td>\n",
       "      <td>27.7</td>\n",
       "      <td>76.2</td>\n",
       "      <td>30.6</td>\n",
       "      <td>58.3</td>\n",
       "      <td>33.7</td>\n",
       "      <td>35.7</td>\n",
       "      <td>...</td>\n",
       "      <td>65.4</td>\n",
       "      <td>29.8</td>\n",
       "      <td>82.3</td>\n",
       "      <td>49.3</td>\n",
       "      <td>25.2</td>\n",
       "      <td>72.2</td>\n",
       "      <td>55.8</td>\n",
       "      <td>35.9</td>\n",
       "      <td>white</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>WE/5</th>\n",
       "      <td>23.6</td>\n",
       "      <td>36.0</td>\n",
       "      <td>29.4</td>\n",
       "      <td>60.6</td>\n",
       "      <td>85.8</td>\n",
       "      <td>58.8</td>\n",
       "      <td>6.8</td>\n",
       "      <td>68.0</td>\n",
       "      <td>37.4</td>\n",
       "      <td>13.3</td>\n",
       "      <td>...</td>\n",
       "      <td>78.3</td>\n",
       "      <td>56.3</td>\n",
       "      <td>36.9</td>\n",
       "      <td>34.9</td>\n",
       "      <td>30.3</td>\n",
       "      <td>52.4</td>\n",
       "      <td>39.6</td>\n",
       "      <td>75.1</td>\n",
       "      <td>black</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 402 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "      playful,serious  shy,bold  cheery,sorrowful  masculine,feminine  \\\n",
       "id                                                                      \n",
       "WE/1             36.4      98.8              72.3                47.3   \n",
       "WE/2             57.4      89.2              64.7                13.8   \n",
       "WE/3             26.8      57.4               9.1                92.1   \n",
       "WE/4             56.9      74.4              28.2                58.4   \n",
       "WE/5             23.6      36.0              29.4                60.6   \n",
       "\n",
       "      charming,awkward  lewd,tasteful  intellectual,physical  strict,lenient  \\\n",
       "id                                                                             \n",
       "WE/1              24.2           29.3                   75.4            64.7   \n",
       "WE/2              19.2           45.0                   67.5            38.9   \n",
       "WE/3              27.1           66.8                   44.0            74.3   \n",
       "WE/4              27.7           76.2                   30.6            58.3   \n",
       "WE/5              85.8           58.8                    6.8            68.0   \n",
       "\n",
       "      refined,rugged  trusting,suspicious  ...  picky,always down  \\\n",
       "id                                         ...                      \n",
       "WE/1            82.1                 57.2  ...               60.4   \n",
       "WE/2            80.4                 54.2  ...               73.9   \n",
       "WE/3            34.0                 12.3  ...               71.4   \n",
       "WE/4            33.7                 35.7  ...               65.4   \n",
       "WE/5            37.4                 13.3  ...               78.3   \n",
       "\n",
       "      proactive,reactive  prudish,flirtatious  innocent,jaded  \\\n",
       "id                                                              \n",
       "WE/1                74.0                 83.9            78.7   \n",
       "WE/2                65.0                 93.3            81.0   \n",
       "WE/3                44.9                 76.8            31.4   \n",
       "WE/4                29.8                 82.3            49.3   \n",
       "WE/5                56.3                 36.9            34.9   \n",
       "\n",
       "      touchy-feely,distant  muddy,washed  quirky,predictable  \\\n",
       "id                                                             \n",
       "WE/1                  70.9          33.2                31.3   \n",
       "WE/2                  66.0          28.1                41.7   \n",
       "WE/3                  13.3          84.4                12.7   \n",
       "WE/4                  25.2          72.2                55.8   \n",
       "WE/5                  30.3          52.4                39.6   \n",
       "\n",
       "      never cries,often crying  race.label  gender.label  \n",
       "id                                                        \n",
       "WE/1                      31.8       white             0  \n",
       "WE/2                      27.5       white             1  \n",
       "WE/3                      72.8       white             0  \n",
       "WE/4                      35.9       white             0  \n",
       "WE/5                      75.1       black             1  \n",
       "\n",
       "[5 rows x 402 columns]"
      ]
     },
     "execution_count": 63,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df = pd.read_csv('data/opendata-with-labels.csv', index_col=0).fillna(0)\n",
    "merged_df['gender.label'] = merged_df['gender.label'].apply(lambda label: 1 if label == 'male' else 0).apply(pd.to_numeric)\n",
    "\n",
    "merged_df.head(5)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 64,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(1462, 402)"
      ]
     },
     "execution_count": 64,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df.shape"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "##### Standardize index names"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 65,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Total Dual Traits: 400\n"
     ]
    }
   ],
   "source": [
    "traits_columns = list(filter(lambda col: ',' in col, merged_df.columns.values))\n",
    "\n",
    "print('Total Dual Traits:', len(traits_columns))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 66,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.model_selection import KFold, train_test_split\n",
    "from sklearn.metrics import mean_squared_error, accuracy_score\n",
    "from datetime import datetime\n",
    "\n",
    "def evaluate_classifier(label, model, classifier, X, y):\n",
    "    \n",
    "    start = datetime.now().timestamp()\n",
    "    \n",
    "    k_fold = KFold(n_splits=5)\n",
    "    \n",
    "    total_accuracy = 0\n",
    "    total_tries = 0\n",
    "    \n",
    "    for train_index, test_index in k_fold.split(X):\n",
    "        X_train, X_test = X[train_index], X[test_index]\n",
    "        y_train, y_test = y[train_index], y[test_index]\n",
    "        \n",
    "        classifier.fit(X_train, y_train)\n",
    "\n",
    "        y_pred = classifier.predict(X_test)\n",
    "        accuracy = accuracy_score(y_test, y_pred, normalize=True)\n",
    "\n",
    "        total_accuracy += accuracy\n",
    "        total_tries += 1\n",
    "        \n",
    "    end = datetime.now().timestamp()\n",
    "    print(f'{label} Accuracy using {model}:', round(total_accuracy / total_tries, 4), ' - seconds taken:', round(end-start, 2))\n",
    "\n",
    "\n",
    "def test_classifier(label, model, classifier, X, y, test_size=0.25):\n",
    "    \n",
    "    start = datetime.now().timestamp()\n",
    "\n",
    "    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=test_size)\n",
    "    \n",
    "    classifier.fit(X_train, y_train)\n",
    "\n",
    "    y_pred = classifier.predict(X_test)\n",
    "    accuracy = accuracy_score(y_test, y_pred, normalize=True)\n",
    "\n",
    "    end = datetime.now().timestamp()\n",
    "    print(f'Final {label} Accuracy using {model}:', round(accuracy, 4), ' - seconds taken:', round(end-start, 2))\n",
    "\n",
    "    \n",
    "\n"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict the Gender"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 67,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Gender Accuracy using AdaBoost: 0.9542  - seconds taken: 9.27\n",
      "Gender Accuracy using Decision Tree, vanilla : 0.8618  - seconds taken: 1.62\n",
      "Gender Accuracy using KNN - 3: 0.8488  - seconds taken: 0.97\n",
      "Gender Accuracy using Logistic Regression L2: 0.9822  - seconds taken: 2.15\n",
      "Gender Accuracy using Multi-layer Perceptron Adam (64): 0.9487  - seconds taken: 15.18\n",
      "Gender Accuracy using Random Forest 2^8 estimators, Gini: 0.9199  - seconds taken: 11.81\n",
      "Gender Accuracy using Random Forest 2^8 estimators, Entropy: 0.9213  - seconds taken: 17.49\n",
      "Gender Accuracy using Ridge Classifier, alpha 0.1: 0.9815  - seconds taken: 0.12\n",
      "Gender Accuracy using SVM Poly (3): 0.9849  - seconds taken: 0.75\n",
      "Gender Accuracy using SVM Poly (5): 0.9863  - seconds taken: 0.76\n",
      "Final Gender Accuracy using Logistic Regression L2: 0.9876  - seconds taken: 0.24\n",
      "Final Gender Accuracy using SVM Poly (3): 0.9917  - seconds taken: 0.13\n",
      "Final Gender Accuracy using SVM Poly (5): 0.9938  - seconds taken: 0.13\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LassoCV, Ridge, RidgeClassifier, LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier\n",
    "from sklearn.model_selection import KFold, train_test_split\n",
    "from sklearn.metrics import mean_squared_error, accuracy_score\n",
    "from sklearn.ensemble import RandomForestClassifier, AdaBoostClassifier\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "from sklearn.neural_network import MLPClassifier\n",
    "from sklearn.naive_bayes import GaussianNB\n",
    "from sklearn import svm\n",
    "\n",
    "# For gender prediction, we remove the 'masculine' label\n",
    "gender_columns = list(filter(lambda trait: 'masculine' not in trait, traits_columns))\n",
    "\n",
    "X = merged_df[gender_columns].values\n",
    "y = merged_df['gender.label'].values\n",
    "\n",
    "\n",
    "evaluate_classifier('Gender', 'AdaBoost', AdaBoostClassifier(), X, y)\n",
    "evaluate_classifier('Gender', 'Decision Tree, vanilla ', DecisionTreeClassifier(), X, y)\n",
    "evaluate_classifier('Gender', 'KNN - 3', KNeighborsClassifier(3), X, y)\n",
    "evaluate_classifier('Gender', 'Logistic Regression L2', LogisticRegression(penalty='l2', max_iter=5000), X, y)\n",
    "evaluate_classifier('Gender', 'Multi-layer Perceptron Adam (64)', MLPClassifier(solver='adam', alpha=0.0001, hidden_layer_sizes=(64,), max_iter=1000), X, y)\n",
    "evaluate_classifier('Gender', 'Random Forest 2^8 estimators, Gini', RandomForestClassifier(n_estimators=2**8, criterion='gini'), X, y)\n",
    "evaluate_classifier('Gender', 'Random Forest 2^8 estimators, Entropy', RandomForestClassifier(n_estimators=2**8, criterion='entropy'), X, y)\n",
    "evaluate_classifier('Gender', 'Ridge Classifier, alpha 0.1', RidgeClassifier(alpha=0.1), X, y)\n",
    "evaluate_classifier('Gender', 'SVM Poly (3)', svm.SVC(kernel='poly', degree=3, C=1.0), X, y)\n",
    "evaluate_classifier('Gender', 'SVM Poly (5)', svm.SVC(kernel='poly', degree=5, C=1.0), X, y)\n",
    "\n",
    "test_classifier('Gender', 'Logistic Regression L2', LogisticRegression(penalty='l2', max_iter=5000), X, y, test_size=0.33)\n",
    "test_classifier('Gender', 'SVM Poly (3)', svm.SVC(kernel='poly', degree=3, C=1.0), X, y, test_size=0.33)\n",
    "test_classifier('Gender', 'SVM Poly (5)', svm.SVC(kernel='poly', degree=5, C=1.0), X, y, test_size=0.33)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1    889\n",
       "0    573\n",
       "Name: gender.label, dtype: int64"
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df['gender.label'].value_counts()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Predict the Race"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 61,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Race Accuracy using Ridge Classifier, alpha 0.1 SVD: 0.8824  - seconds taken: 2.44\n",
      "Race Accuracy using Ridge Classifier, alpha 0.1 Cholesky: 0.8824  - seconds taken: 0.14\n",
      "Race Accuracy using Ridge Classifier, alpha 0.1 lsqr: 0.8851  - seconds taken: 0.49\n",
      "Race Accuracy using Ridge Classifier, alpha 0.1 sag: 0.8837  - seconds taken: 21.01\n",
      "Race Accuracy using Ridge Classifier, alpha 0.1 saga: 0.8851  - seconds taken: 30.97\n"
     ]
    }
   ],
   "source": [
    "from sklearn.linear_model import LassoCV, Ridge, RidgeClassifier, LogisticRegression\n",
    "from sklearn.tree import DecisionTreeClassifier, plot_tree\n",
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.model_selection import KFold, train_test_split\n",
    "from sklearn.metrics import mean_squared_error, accuracy_score\n",
    "from sklearn.preprocessing import LabelEncoder\n",
    "from sklearn.neighbors import KNeighborsClassifier\n",
    "\n",
    "# For gender prediction, we remove the 'masculine' label\n",
    "#race_columns = list(traits_columns)\n",
    "race_columns = list(filter(lambda trait: 'western' not in trait, traits_columns))\n",
    "race_columns.append('gender.label')\n",
    "\n",
    "y_encoder = LabelEncoder()\n",
    "\n",
    "X = merged_df[race_columns].values\n",
    "y = y_encoder.fit_transform(merged_df['race.label'])\n",
    "\n",
    "evaluate_classifier('Race', 'AdaBoost', AdaBoostClassifier(), X, y)\n",
    "evaluate_classifier('Race', 'Decision Tree, vanilla ', DecisionTreeClassifier(), X, y)\n",
    "evaluate_classifier('Race', 'KNN - 3', KNeighborsClassifier(3), X, y)\n",
    "evaluate_classifier('Race', 'Logistic Regression L2, 10000 iterations', LogisticRegression(penalty='l2', max_iter=10000), X, y)\n",
    "evaluate_classifier('Race', 'Multi-layer Perceptron Adam (64)', MLPClassifier(solver='adam', alpha=0.0001, hidden_layer_sizes=(64,), max_iter=1000), X, y)\n",
    "evaluate_classifier('Race', 'Random Forest 2^8 estimators, Gini', RandomForestClassifier(n_estimators=2**8, criterion='gini'), X, y)\n",
    "evaluate_classifier('Race', 'Random Forest 2^8 estimators, Entropy', RandomForestClassifier(n_estimators=2**8, criterion='entropy'), X, y)\n",
    "evaluate_classifier('Race', 'Ridge Classifier, alpha 0.1 SVD', RidgeClassifier(alpha=0.1, solver='svd'), X, y)\n",
    "evaluate_classifier('Race', 'SVM Poly (3)', svm.SVC(kernel='poly', degree=3, C=1.0), X, y)\n",
    "evaluate_classifier('Race', 'SVM Poly (5)', svm.SVC(kernel='poly', degree=5, C=1.0), X, y)\n",
    "\n",
    "test_classifier('Race', 'Random Forest 2^8 estimators, Entropy', RandomForestClassifier(n_estimators=2**8, criterion='entropy'), X, y, test_size=0.33)\n",
    "test_classifier('Race', 'Ridge Classifier, alpha 0.1', RidgeClassifier(alpha=0.1), X, y, test_size=0.33)\n",
    "#test_classifier('Race', 'SVM Poly (3)', svm.SVC(kernel='poly', degree=3, C=1.0), X, y, test_size=0.33)\n",
    "#test_classifier('Race', 'SVM Poly (5)', svm.SVC(kernel='poly', degree=5, C=1.0), X, y, test_size=0.33)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 85,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1440x720 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "from sklearn import tree\n",
    "from matplotlib import pyplot as plt\n",
    "\n",
    "decision_tree_classifier = DecisionTreeClassifier(criterion='gini')  \n",
    "decision_tree_classifier.fit(X, y)\n",
    "\n",
    "\n",
    "\n",
    "fig = plt.figure(figsize=(20,10))\n",
    "_ = tree.plot_tree(decision_tree_classifier, \n",
    "                   feature_names=race_columns,  \n",
    "                   class_names=y_encoder.classes_,\n",
    "                   filled=True,\n",
    "                   label='all')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "0.8180574555403557"
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df['race.label'].value_counts()['white'] / len(merged_df['race.label'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "white              1196\n",
       "black               136\n",
       "asian                60\n",
       "hispanic             54\n",
       "mideast               7\n",
       "indian                7\n",
       "native american       2\n",
       "Name: race.label, dtype: int64"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "merged_df['race.label'].value_counts()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1462"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(merged_df['race.label'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}