mananai/user_classifier_full.ipynb

## user_classifier_full.ipynb
{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import sqlite3\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "3.7.7\n"
     ]
    }
   ],
   "source": [
    "from platform import python_version\n",
    "\n",
    "print(python_version())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<sqlite3.Cursor at 0x12199dbb20>"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "from pathlib import Path\n",
    "db_dir = str(Path.home()) + '/sqlitedb/'\n",
    "classifier_file = db_dir + 'TwitterUserClassifier.db'\n",
    "friend_file = db_dir + 'TwitterFriends.db'\n",
    "conn = sqlite3.connect(classifier_file)\n",
    "conn.execute('ATTACH DATABASE \"' + friend_file + '\" AS friends')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "User class1 shape: \n",
      "(3038, 2)\n",
      "User class0 shape: \n",
      "(5261, 2)\n"
     ]
    }
   ],
   "source": [
    "#dfUserClass1 =  pd.read_sql_query('select id, 1 class from user_class1 order by random();', conn)\n",
    "#dfUserClass0 =  pd.read_sql_query('select id, 0 class from user_class0 order by random();', conn)\n",
    "dfUserClass1 =  pd.read_sql_query('select distinct u.id, 1 class '\n",
    "                                  'from user_class1 u '\n",
    "                                  'join friends.user_friend uf '\n",
    "                                  'on u.id=uf.user_id and uf.friend_id is not null '\n",
    "                                  'order by random();',\n",
    "                                  conn)\n",
    "dfUserClass0 =  pd.read_sql_query('select distinct u.id, 0 class '\n",
    "                                  'from user_class0 u '\n",
    "                                  'join friends.user_friend uf '\n",
    "                                  'on u.id=uf.user_id and uf.friend_id is not null '\n",
    "                                  'order by random();', conn)\n",
    "print(\"User class1 shape: \")\n",
    "print(dfUserClass1.shape)\n",
    "print(\"User class0 shape: \")\n",
    "print(dfUserClass0.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "dfUserClass1 = dfUserClass1.head(3000)\n",
    "dfUserClass0 = dfUserClass0.head(5000)\n",
    "dfUserClassAll = pd.concat([dfUserClass1, dfUserClass0])\n",
    "dfUserClassAll = dfUserClassAll.sample(frac=1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "User class all shape: \n",
      "(8000, 2)\n",
      "Data Set Train shape: \n",
      "(5600, 2)\n",
      "Data Set Test shape: \n",
      "(2400, 2)\n"
     ]
    }
   ],
   "source": [
    "splitPos = int(dfUserClassAll.shape[0]*0.70)\n",
    "dfTrain = dfUserClassAll.iloc[0:splitPos]\n",
    "dfTest = dfUserClassAll.iloc[splitPos:]\n",
    "print(\"User class all shape: \")\n",
    "print(dfUserClassAll.shape)\n",
    "print(\"Data Set Train shape: \")\n",
    "print(dfTrain.shape)\n",
    "print(\"Data Set Test shape: \")\n",
    "print(dfTest.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 3417929 entries, 0 to 3417928\n",
      "Data columns (total 2 columns):\n",
      " #   Column     Dtype\n",
      "---  ------     -----\n",
      " 0   user_id    int64\n",
      " 1   friend_id  int64\n",
      "dtypes: int64(2)\n",
      "memory usage: 52.2 MB\n"
     ]
    }
   ],
   "source": [
    "dfUserFriend=pd.read_sql_query('select user_id, friend_id '\n",
    "                               'from friends.user_friend '\n",
    "                               'where friend_id is not null;', conn)\n",
    "dfUserFriend.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "<class 'pandas.core.frame.DataFrame'>\n",
      "RangeIndex: 709151 entries, 0 to 709150\n",
      "Data columns (total 4 columns):\n",
      " #   Column           Non-Null Count   Dtype \n",
      "---  ------           --------------   ----- \n",
      " 0   id               709151 non-null  int64 \n",
      " 1   screen_name      709151 non-null  object\n",
      " 2   name             709151 non-null  object\n",
      " 3   followers_count  709151 non-null  int64 \n",
      "dtypes: int64(2), object(2)\n",
      "memory usage: 21.6+ MB\n"
     ]
    }
   ],
   "source": [
    "dfAllUsers = pd.read_sql_query('select u.id, u.screen_name, u.name, u.followers_count '\n",
    "                           'from friends.user u;', conn)\n",
    "dfAllUsers.set_index(['id'])\n",
    "dfAllUsers.info()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [],
   "source": [
    "def buildTopUser(_df_user_friend, _df_user, _numTopUsers):\n",
    "    _i1 = _df_user_friend.set_index(['user_id']).index\n",
    "    _i2 = _df_user.set_index(['id']).index\n",
    "    _dfTopUser = _df_user_friend[_i1.isin(_i2)]\n",
    "    _dfTopUser = _dfTopUser.groupby(['friend_id'])['user_id'].count().reset_index(name='count').sort_values(['count'], ascending=False)[:_numTopUsers]\n",
    "    #_dfTopUser = pd.merge(_dfTopUser, _df_all_users, left_on='friend_id', right_on='id')\n",
    "    return _dfTopUser"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "def buildFeatureVectors(_df_user_friend, _df_user, _df_top_user):\n",
    "    _i1 = _df_user_friend.set_index('user_id').index\n",
    "    _i2 = _df_user.set_index('id').index\n",
    "    _dfUserFriend2 = _df_user_friend[_i1.isin(_i2)]\n",
    "    _i3 = _dfUserFriend2.set_index('friend_id').index\n",
    "    _i4 = _df_top_user.set_index('friend_id').index\n",
    "    _dfUserFriend3 = _dfUserFriend2[_i3.isin(_i4)]\n",
    "    \n",
    "    _feature_vectors = pd.crosstab(index=_dfUserFriend3['user_id'], columns=_dfUserFriend3['friend_id'], dropna=False)\n",
    "    _feature_vectors = pd.merge(_feature_vectors, _df_user, left_on='user_id', right_on='id')\n",
    "    return _feature_vectors\n",
    "    "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "from sklearn.ensemble import RandomForestClassifier\n",
    "from sklearn.metrics import classification_report, confusion_matrix, accuracy_score\n",
    "from sklearn.metrics import roc_auc_score"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "numTopUser = 500\n",
    "dfTopUser = buildTopUser(dfUserFriend, dfTrain, numTopUser)\n",
    "numTopUser=dfTopUser.shape[0]\n",
    "feature_vectors_train = buildFeatureVectors(dfUserFriend, dfTrain, dfTopUser)\n",
    "feature_vectors_test = buildFeatureVectors(dfUserFriend, dfTest, dfTopUser)\n",
    "#Rearrange columns on the test\n",
    "feature_vectors_test = feature_vectors_test[feature_vectors_train.columns]\n",
    "X_train = feature_vectors_train.iloc[:, :-2].values\n",
    "y_train = feature_vectors_train.iloc[:, (numTopUser+1)].values\n",
    "X_test = feature_vectors_test.iloc[:, :-2].values\n",
    "y_test = feature_vectors_test.iloc[:, (numTopUser+1)].values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "RandomForestClassifier()"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "classifier = RandomForestClassifier()\n",
    "classifier.fit(X_train, y_train)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "ROC AUC:  0.9346447723041871\n",
      "Classification Report:\n",
      "              precision    recall  f1-score   support\n",
      "\n",
      "           0       0.88      0.96      0.92      1395\n",
      "           1       0.92      0.79      0.85       903\n",
      "\n",
      "    accuracy                           0.89      2298\n",
      "   macro avg       0.90      0.87      0.88      2298\n",
      "weighted avg       0.89      0.89      0.89      2298\n",
      "\n",
      "Accuracy: 0.8920800696257616\n",
      "Confusion Matrix:\n",
      "[[1336   59]\n",
      " [ 189  714]]\n"
     ]
    }
   ],
   "source": [
    "y_pred = classifier.predict(X_test)\n",
    "y_prob = classifier.predict_proba(X_test)[:, 1]\n",
    "print(\"ROC AUC: \", roc_auc_score(y_test, y_prob))\n",
    "print(\"Classification Report:\",)\n",
    "print (classification_report(y_test, y_pred))\n",
    "print(\"Accuracy:\", accuracy_score(y_test,y_pred))\n",
    "print(\"Confusion Matrix:\")\n",
    "print(confusion_matrix(y_test, y_pred))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>feature</th>\n",
       "      <th>importance</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>496</th>\n",
       "      <td>1301760719906512896</td>\n",
       "      <td>0.066097</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>202</th>\n",
       "      <td>199992361</td>\n",
       "      <td>0.051649</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>142</th>\n",
       "      <td>92531902</td>\n",
       "      <td>0.040005</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>69</th>\n",
       "      <td>50916567</td>\n",
       "      <td>0.039931</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>487</th>\n",
       "      <td>1269106344528736256</td>\n",
       "      <td>0.028700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>278</th>\n",
       "      <td>953670151</td>\n",
       "      <td>0.000104</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>125</th>\n",
       "      <td>78871747</td>\n",
       "      <td>0.000100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>348</th>\n",
       "      <td>3968042414</td>\n",
       "      <td>0.000100</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>206</th>\n",
       "      <td>213612447</td>\n",
       "      <td>0.000078</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>270</th>\n",
       "      <td>830434249</td>\n",
       "      <td>0.000030</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>500 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "                 feature  importance\n",
       "496  1301760719906512896    0.066097\n",
       "202            199992361    0.051649\n",
       "142             92531902    0.040005\n",
       "69              50916567    0.039931\n",
       "487  1269106344528736256    0.028700\n",
       "..                   ...         ...\n",
       "278            953670151    0.000104\n",
       "125             78871747    0.000100\n",
       "348           3968042414    0.000100\n",
       "206            213612447    0.000078\n",
       "270            830434249    0.000030\n",
       "\n",
       "[500 rows x 2 columns]"
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "fi = pd.DataFrame({'feature': list(feature_vectors_train.columns[:-2]),\n",
    "                   'importance': classifier.feature_importances_}).\\\n",
    "                    sort_values('importance', ascending = False)\n",
    "fi"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x1226b1f110>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x504 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "fi[:40].plot.bar(x='feature', y='importance', figsize=(14,7))"
   ]
  },
  {
   "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": "3.7.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}