COLABORATIVE_FILTERING.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.6/importlib/_bootstrap.py:219: RuntimeWarning: compiletime version 3.5 of module 'tensorflow.python.framework.fast_tensor_util' does not match runtime version 3.6\n",
      "  return f(*args, **kwds)\n",
      "/anaconda3/lib/python3.6/site-packages/h5py/__init__.py:36: FutureWarning: Conversion of the second argument of issubdtype from `float` to `np.floating` is deprecated. In future, it will be treated as `np.float64 == np.dtype(float).type`.\n",
      "  from ._conv import register_converters as _register_converters\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "\n",
    "#imports\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import tensorflow as tf\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Get Data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_creation_df = pd.read_csv(f'data/all_label_creations.csv')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Some properties are duplicated in the CSV...\n",
    "label_creation_df.drop_duplicates(inplace=True)\n",
    "# Sort revisions by properties and timestamps\n",
    "label_creation_df.sort_values([\"property\",\"timestamp\"],inplace=True)\n",
    "# Delete properties with NaN values\n",
    "label_creation_df = label_creation_df.loc[~label_creation_df.language.isna()]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "label_creation_df['rating'] = 1"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>timestamp</th>\n",
       "      <th>property</th>\n",
       "      <th>language</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>89086</th>\n",
       "      <td>2013-10-21T12:00:34Z</td>\n",
       "      <td>P969</td>\n",
       "      <td>nl</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  timestamp property language  rating\n",
       "89086  2013-10-21T12:00:34Z     P969       nl       1"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_creation_df.sample()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "PROPERTIES: 5126 LANGUAGES: 234\n"
     ]
    }
   ],
   "source": [
    "num_languages = label_creation_df.language.nunique()\n",
    "num_properties = label_creation_df.property.nunique()\n",
    "\n",
    "print(\"PROPERTIES: {} LANGUAGES: {}\".format(num_properties, num_languages))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Normalize in [0, 1]\n",
    "\n",
    "#r = df['rating'].values.astype(float)\n",
    "#min_max_scaler = preprocessing.MinMaxScaler()\n",
    "#x_scaled = min_max_scaler.fit_transform(r.reshape(-1,1))\n",
    "#df_normalized = pd.DataFrame(x_scaled)\n",
    "#df['rating'] = df_normalized"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>timestamp</th>\n",
       "      <th>property</th>\n",
       "      <th>language</th>\n",
       "      <th>rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>66867</th>\n",
       "      <td>2013-10-12T17:13:46Z</td>\n",
       "      <td>P463</td>\n",
       "      <td>lb</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                  timestamp property language  rating\n",
       "66867  2013-10-12T17:13:46Z     P463       lb       1"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "label_creation_df.sample()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.6/site-packages/ipykernel_launcher.py:11: FutureWarning: Method .as_matrix will be removed in a future version. Use .values instead.\n",
      "  # This is added back by InteractiveShellApp.init_path()\n"
     ]
    }
   ],
   "source": [
    "# Convert DataFrame in user-item matrix\n",
    "\n",
    "matrix = label_creation_df.pivot_table(index='property', columns='language', values='rating')\n",
    "matrix.fillna(0, inplace=True)\n",
    "\n",
    "# Users and items ordered as they are in matrix\n",
    "\n",
    "properties = matrix.index.tolist()\n",
    "languages = matrix.columns.tolist()\n",
    "\n",
    "matrix = matrix.as_matrix()\n",
    "#matrix = matrix.values()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Network Parameters\n",
    "\n",
    "num_input = num_languages\n",
    "num_hidden_1 = 30\n",
    "num_hidden_2 = 10\n",
    "\n",
    "X = tf.placeholder(tf.float64, [None, num_input])\n",
    "\n",
    "weights = {\n",
    "    'encoder_h1': tf.Variable(tf.random_normal([num_input, num_hidden_1], dtype=tf.float64)),\n",
    "    'encoder_h2': tf.Variable(tf.random_normal([num_hidden_1, num_hidden_2], dtype=tf.float64)),\n",
    "    'decoder_h1': tf.Variable(tf.random_normal([num_hidden_2, num_hidden_1], dtype=tf.float64)),\n",
    "    'decoder_h2': tf.Variable(tf.random_normal([num_hidden_1, num_input], dtype=tf.float64)),\n",
    "}\n",
    "\n",
    "biases = {\n",
    "    'encoder_b1': tf.Variable(tf.random_normal([num_hidden_1], dtype=tf.float64)),\n",
    "    'encoder_b2': tf.Variable(tf.random_normal([num_hidden_2], dtype=tf.float64)),\n",
    "    'decoder_b1': tf.Variable(tf.random_normal([num_hidden_1], dtype=tf.float64)),\n",
    "    'decoder_b2': tf.Variable(tf.random_normal([num_input], dtype=tf.float64)),\n",
    "}"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Building the encoder\n",
    "\n",
    "def encoder(x):\n",
    "    # Encoder Hidden layer with sigmoid activation #1\n",
    "    layer_1 = tf.nn.sigmoid(tf.add(tf.matmul(x, weights['encoder_h1']), biases['encoder_b1']))\n",
    "    # Encoder Hidden layer with sigmoid activation #2\n",
    "    layer_2 = tf.nn.sigmoid(tf.add(tf.matmul(layer_1, weights['encoder_h2']), biases['encoder_b2']))\n",
    "    return layer_2\n",
    "\n",
    "\n",
    "# Building the decoder\n",
    "\n",
    "def decoder(x):\n",
    "    # Decoder Hidden layer with sigmoid activation #1\n",
    "    layer_1 = tf.nn.sigmoid(tf.add(tf.matmul(x, weights['decoder_h1']), biases['decoder_b1']))\n",
    "    # Decoder Hidden layer with sigmoid activation #2\n",
    "    layer_2 = tf.nn.sigmoid(tf.add(tf.matmul(layer_1, weights['decoder_h2']), biases['decoder_b2']))\n",
    "    return layer_2\n",
    "\n",
    "\n",
    "# Construct model\n",
    "\n",
    "encoder_op = encoder(X)\n",
    "decoder_op = decoder(encoder_op)\n",
    "\n",
    "\n",
    "# Prediction\n",
    "\n",
    "y_pred = decoder_op\n",
    "\n",
    "\n",
    "# Targets are the input data.\n",
    "\n",
    "y_true = X"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Define loss and optimizer, minimize the squared error\n",
    "\n",
    "loss = tf.losses.mean_squared_error(y_true, y_pred)\n",
    "optimizer = tf.train.RMSPropOptimizer(0.03).minimize(loss)\n",
    "\n",
    "predictions = pd.DataFrame()\n",
    "\n",
    "# Define evaluation metrics\n",
    "\n",
    "eval_x = tf.placeholder(tf.int32, )\n",
    "eval_y = tf.placeholder(tf.int32, )\n",
    "pre, pre_op = tf.metrics.precision(labels=eval_x, predictions=eval_y)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Initialize the variables (i.e. assign their default value)\n",
    "\n",
    "init = tf.global_variables_initializer()\n",
    "local_init = tf.local_variables_initializer()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Predictions...\n"
     ]
    }
   ],
   "source": [
    "with tf.Session() as session:\n",
    "    epochs = 100\n",
    "    batch_size = 25\n",
    "\n",
    "    session.run(init)\n",
    "    session.run(local_init)\n",
    "\n",
    "    num_batches = int(matrix.shape[0] / batch_size)\n",
    "    matrix = np.array_split(matrix, num_batches)\n",
    "\n",
    "    for i in range(epochs):\n",
    "\n",
    "        avg_cost = 0\n",
    "\n",
    "        for batch in matrix:\n",
    "            _, l = session.run([optimizer, loss], feed_dict={X: batch})\n",
    "            avg_cost += l\n",
    "\n",
    "        avg_cost /= num_batches\n",
    "\n",
    "        #print(\"Epoch: {} Loss: {}\".format(i + 1, avg_cost))\n",
    "\n",
    "    print(\"Predictions...\")\n",
    "\n",
    "    matrix_copy = np.concatenate(matrix, axis=0)\n",
    "\n",
    "    preds = session.run(decoder_op, feed_dict={X: matrix_copy})\n",
    "\n",
    "    predictions = predictions.append(pd.DataFrame(preds))\n",
    "\n",
    "    predictions = predictions.stack().reset_index(name='rating')\n",
    "    predictions.columns = ['property', 'language','rating']\n",
    "    predictions['property'] = predictions['property'].map(lambda value: properties[value])\n",
    "    predictions['language'] = predictions['language'].map(lambda value: languages[value])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [],
   "source": [
    "predictions = predictions.sort_values(by=['rating'],ascending=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "rrr = label_creation_df.pivot_table(index='property', columns='language', values='rating')\n",
    "rrr.fillna(0, inplace=True)\n",
    "rrr= pd.DataFrame(rrr.unstack())\n",
    "rrr.reset_index(inplace=True)\n",
    "rrr.columns = [ 'language','property','actual_rating']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "aaa = pd.merge(predictions,rrr,how='inner',left_on=['property','language'],right_on=['property','language'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Show the prediction distribution"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.6/site-packages/matplotlib/axes/_axes.py:6462: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x113fadda0>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,8))\n",
    "sns.distplot(aaa.rating)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.6/site-packages/matplotlib/axes/_axes.py:6462: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x113fadb00>"
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,8))\n",
    "sns.distplot(aaa.loc[aaa.actual_rating == 1,'rating'])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/anaconda3/lib/python3.6/site-packages/matplotlib/axes/_axes.py:6462: UserWarning: The 'normed' kwarg is deprecated, and has been replaced by the 'density' kwarg.\n",
      "  warnings.warn(\"The 'normed' kwarg is deprecated, and has been \"\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x10e515da0>"
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 864x576 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.figure(figsize=(12,8))\n",
    "sns.distplot(aaa.loc[aaa.actual_rating == 0,'rating'])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Show the prediction"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "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>property</th>\n",
       "      <th>language</th>\n",
       "      <th>rating</th>\n",
       "      <th>actual_rating</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>519</th>\n",
       "      <td>P4010</td>\n",
       "      <td>de</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2073</th>\n",
       "      <td>P2335</td>\n",
       "      <td>de</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2118</th>\n",
       "      <td>P2336</td>\n",
       "      <td>de</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2258</th>\n",
       "      <td>P3047</td>\n",
       "      <td>es</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2263</th>\n",
       "      <td>P4227</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3191</th>\n",
       "      <td>P3069</td>\n",
       "      <td>de</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3522</th>\n",
       "      <td>P2225</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3703</th>\n",
       "      <td>P913</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4903</th>\n",
       "      <td>P1900</td>\n",
       "      <td>uk</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4904</th>\n",
       "      <td>P2005</td>\n",
       "      <td>uk</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5037</th>\n",
       "      <td>P1277</td>\n",
       "      <td>ca</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5041</th>\n",
       "      <td>P2355</td>\n",
       "      <td>de</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5296</th>\n",
       "      <td>P3212</td>\n",
       "      <td>uk</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5302</th>\n",
       "      <td>P1438</td>\n",
       "      <td>es</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5865</th>\n",
       "      <td>P4584</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6206</th>\n",
       "      <td>P556</td>\n",
       "      <td>ia</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6657</th>\n",
       "      <td>P974</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6832</th>\n",
       "      <td>P787</td>\n",
       "      <td>es</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7123</th>\n",
       "      <td>P2671</td>\n",
       "      <td>pl</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7518</th>\n",
       "      <td>P532</td>\n",
       "      <td>pl</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7698</th>\n",
       "      <td>P1102</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7783</th>\n",
       "      <td>P1148</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7903</th>\n",
       "      <td>P3189</td>\n",
       "      <td>mk</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8219</th>\n",
       "      <td>P945</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8927</th>\n",
       "      <td>P3300</td>\n",
       "      <td>pl</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8970</th>\n",
       "      <td>P1881</td>\n",
       "      <td>zh-hans</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9323</th>\n",
       "      <td>P3158</td>\n",
       "      <td>mk</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9361</th>\n",
       "      <td>P2206</td>\n",
       "      <td>es</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10018</th>\n",
       "      <td>P5317</td>\n",
       "      <td>es</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10327</th>\n",
       "      <td>P2987</td>\n",
       "      <td>de</td>\n",
       "      <td>1.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199454</th>\n",
       "      <td>P2483</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199455</th>\n",
       "      <td>P4459</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199456</th>\n",
       "      <td>P5722</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199457</th>\n",
       "      <td>P4222</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199458</th>\n",
       "      <td>P4154</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199459</th>\n",
       "      <td>P5287</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199460</th>\n",
       "      <td>P5675</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199461</th>\n",
       "      <td>P4204</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199462</th>\n",
       "      <td>P5572</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199463</th>\n",
       "      <td>P2409</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199464</th>\n",
       "      <td>P5197</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199465</th>\n",
       "      <td>P2939</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199466</th>\n",
       "      <td>P3343</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199467</th>\n",
       "      <td>P4425</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199468</th>\n",
       "      <td>P3021</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199469</th>\n",
       "      <td>P3548</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199470</th>\n",
       "      <td>P5621</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199471</th>\n",
       "      <td>P5581</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199472</th>\n",
       "      <td>P2966</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199473</th>\n",
       "      <td>P3213</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199474</th>\n",
       "      <td>P4228</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199475</th>\n",
       "      <td>P4533</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199476</th>\n",
       "      <td>P2879</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199477</th>\n",
       "      <td>P4426</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199478</th>\n",
       "      <td>P4153</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199479</th>\n",
       "      <td>P4231</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199480</th>\n",
       "      <td>P5041</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199481</th>\n",
       "      <td>P5042</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199482</th>\n",
       "      <td>P5040</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1199483</th>\n",
       "      <td>P4448</td>\n",
       "      <td>ia</td>\n",
       "      <td>0.0000</td>\n",
       "      <td>0.0000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>1089070 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "        property language  rating  actual_rating\n",
       "519        P4010       de  1.0000         0.0000\n",
       "2073       P2335       de  1.0000         0.0000\n",
       "2118       P2336       de  1.0000         0.0000\n",
       "2258       P3047       es  1.0000         0.0000\n",
       "2263       P4227  zh-hans  1.0000         0.0000\n",
       "3191       P3069       de  1.0000         0.0000\n",
       "3522       P2225  zh-hans  1.0000         0.0000\n",
       "3703        P913  zh-hans  1.0000         0.0000\n",
       "4903       P1900       uk  1.0000         0.0000\n",
       "4904       P2005       uk  1.0000         0.0000\n",
       "5037       P1277       ca  1.0000         0.0000\n",
       "5041       P2355       de  1.0000         0.0000\n",
       "5296       P3212       uk  1.0000         0.0000\n",
       "5302       P1438       es  1.0000         0.0000\n",
       "5865       P4584  zh-hans  1.0000         0.0000\n",
       "6206        P556       ia  1.0000         0.0000\n",
       "6657        P974  zh-hans  1.0000         0.0000\n",
       "6832        P787       es  1.0000         0.0000\n",
       "7123       P2671       pl  1.0000         0.0000\n",
       "7518        P532       pl  1.0000         0.0000\n",
       "7698       P1102  zh-hans  1.0000         0.0000\n",
       "7783       P1148  zh-hans  1.0000         0.0000\n",
       "7903       P3189       mk  1.0000         0.0000\n",
       "8219        P945  zh-hans  1.0000         0.0000\n",
       "8927       P3300       pl  1.0000         0.0000\n",
       "8970       P1881  zh-hans  1.0000         0.0000\n",
       "9323       P3158       mk  1.0000         0.0000\n",
       "9361       P2206       es  1.0000         0.0000\n",
       "10018      P5317       es  1.0000         0.0000\n",
       "10327      P2987       de  1.0000         0.0000\n",
       "...          ...      ...     ...            ...\n",
       "1199454    P2483       ia  0.0000         0.0000\n",
       "1199455    P4459       ia  0.0000         0.0000\n",
       "1199456    P5722       ia  0.0000         0.0000\n",
       "1199457    P4222       ia  0.0000         0.0000\n",
       "1199458    P4154       ia  0.0000         0.0000\n",
       "1199459    P5287       ia  0.0000         0.0000\n",
       "1199460    P5675       ia  0.0000         0.0000\n",
       "1199461    P4204       ia  0.0000         0.0000\n",
       "1199462    P5572       ia  0.0000         0.0000\n",
       "1199463    P2409       ia  0.0000         0.0000\n",
       "1199464    P5197       ia  0.0000         0.0000\n",
       "1199465    P2939       ia  0.0000         0.0000\n",
       "1199466    P3343       ia  0.0000         0.0000\n",
       "1199467    P4425       ia  0.0000         0.0000\n",
       "1199468    P3021       ia  0.0000         0.0000\n",
       "1199469    P3548       ia  0.0000         0.0000\n",
       "1199470    P5621       ia  0.0000         0.0000\n",
       "1199471    P5581       ia  0.0000         0.0000\n",
       "1199472    P2966       ia  0.0000         0.0000\n",
       "1199473    P3213       ia  0.0000         0.0000\n",
       "1199474    P4228       ia  0.0000         0.0000\n",
       "1199475    P4533       ia  0.0000         0.0000\n",
       "1199476    P2879       ia  0.0000         0.0000\n",
       "1199477    P4426       ia  0.0000         0.0000\n",
       "1199478    P4153       ia  0.0000         0.0000\n",
       "1199479    P4231       ia  0.0000         0.0000\n",
       "1199480    P5041       ia  0.0000         0.0000\n",
       "1199481    P5042       ia  0.0000         0.0000\n",
       "1199482    P5040       ia  0.0000         0.0000\n",
       "1199483    P4448       ia  0.0000         0.0000\n",
       "\n",
       "[1089070 rows x 4 columns]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.options.display.float_format = '{:,.4f}'.format\n",
    "aaa.loc[(aaa.actual_rating == 0)\n",
    "        & (aaa.rating < 1)\n",
    "        & (~aaa.language.isin(['fr',]))]"
   ]
  },
  {
   "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": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}