ICM visualization and preprocessing.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "a9d14c83",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import pandas as pd"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "43e5feeb",
   "metadata": {},
   "source": [
    "### Import datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "8cd7027d",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/var/folders/f1/t7pg4vg155x4hjw8rxwkp19w0000gn/T/ipykernel_51835/2485813128.py:3: DtypeWarning: Columns (2) have mixed types. Specify dtype option on import or set low_memory=False.\n",
      "  interactions = pd.read_csv(\"data/interactions_and_impressions.csv\")\n"
     ]
    }
   ],
   "source": [
    "data_ICM_type = pd.read_csv(\"data/data_ICM_type.csv\")\n",
    "data_ICM_length = pd.read_csv(\"data/data_ICM_length.csv\", index_col=False)\n",
    "interactions = pd.read_csv(\"data/interactions_and_impressions.csv\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "3e2909ff",
   "metadata": {},
   "outputs": [],
   "source": [
    "all_items_id = pd.concat([interactions[\"ItemID\"], data_ICM_length[\"item_id\"], data_ICM_type[\"item_id\"]], ignore_index=True).unique()\n",
    "data_ICM_length.drop(columns=[\"feature_id\", \"item_id\"], inplace=True)\n",
    "data_ICM_type.drop(columns=[\"data\",\"item_id\"], inplace=True)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e961f8e0",
   "metadata": {},
   "source": [
    "#### Control null values in the datasets"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "313fb245",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Number of null values in ICM_length: 0\n",
      "Number of null values in ICM_type: 0\n"
     ]
    }
   ],
   "source": [
    "print(\"Number of null values in ICM_length: \"+str(data_ICM_length[\"data\"].isna().sum()))\n",
    "print(\"Number of null values in ICM_type: \"+str(data_ICM_type[\"feature_id\"].isna().sum()))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b1775179",
   "metadata": {},
   "source": [
    "### Visualize ICM_length dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "a66b77ec",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_ICM_length[\"data\"].plot(figsize=(10,4))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "da3d0d06",
   "metadata": {},
   "source": [
    "#### Removing outliers (values > 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "f7fb1da0",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 11,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x400 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_ICM_length.drop(data_ICM_length[data_ICM_length[\"data\"]>1000].index, inplace=True)\n",
    "data_ICM_length[\"data\"].plot(figsize=(10,4))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "17d57edb",
   "metadata": {},
   "source": [
    "#### Visualize sorted values"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "id": "3aa627e8",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<AxesSubplot:>"
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1000x500 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "data_ICM_length_sorted = pd.DataFrame(data_ICM_length.sort_values(by=[\"data\"], ascending=False))\n",
    "data_ICM_length_sorted.index = [x for x in range(1, len(data_ICM_length.data)+1)]\n",
    "data_ICM_length_sorted.plot(figsize=(10,5))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "991aed7a",
   "metadata": {},
   "source": [
    "## One-hot encoding"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "7b4ff2a4",
   "metadata": {},
   "source": [
    "#### Groups division:\n",
    "1 puntata -> \"breve\"\n",
    "\n",
    "tra 2 e 30 -> \"media\"\n",
    "\n",
    "30 e 80 -> \"lunga\"\n",
    "\n",
    "80 e 1000 -> \"molto lunga\"\n",
    "\n",
    "1000 e 10000 -> \"outlier\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ea92ee8f",
   "metadata": {},
   "source": [
    "####  Create column for one-hot encoding (ICM_length)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "a82dcafe",
   "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>dummy_long</th>\n",
       "      <th>dummy_medium</th>\n",
       "      <th>dummy_one_episode</th>\n",
       "      <th>dummy_very_long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</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>23086</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23087</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23088</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23089</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23090</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>23012 rows × 4 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       dummy_long  dummy_medium  dummy_one_episode  dummy_very_long\n",
       "0               0             0                  1                0\n",
       "1               0             0                  1                0\n",
       "2               0             1                  0                0\n",
       "3               0             0                  1                0\n",
       "4               0             0                  1                0\n",
       "...           ...           ...                ...              ...\n",
       "23086           0             0                  1                0\n",
       "23087           0             0                  1                0\n",
       "23088           0             0                  1                0\n",
       "23089           0             0                  1                0\n",
       "23090           0             1                  0                0\n",
       "\n",
       "[23012 rows x 4 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "length = []\n",
    "for index, row in data_ICM_length.iterrows():\n",
    "    if(row[\"data\"] == 1):\n",
    "        length.append(\"one_episode\")\n",
    "    elif(row[\"data\"]>1 and row[\"data\"]<30):\n",
    "        length.append(\"medium\")\n",
    "    elif(row[\"data\"]>=30 and row[\"data\"]<80):\n",
    "        length.append(\"long\")\n",
    "    elif(row[\"data\"]>=80 and row[\"data\"]<1000):\n",
    "        length.append(\"very_long\")\n",
    "    else:\n",
    "        length.append(\"outlier\")\n",
    "\n",
    "data_ICM_length[\"dummy\"] = length\n",
    "one_hot = pd.get_dummies(data_ICM_length.dummy, prefix='dummy')\n",
    "one_hot"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "id": "c6316c14",
   "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>dummy_long</th>\n",
       "      <th>dummy_medium</th>\n",
       "      <th>dummy_one_episode</th>\n",
       "      <th>dummy_outlier</th>\n",
       "      <th>dummy_very_long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23086</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23087</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23088</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23089</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23090</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>23091 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       dummy_long  dummy_medium  dummy_one_episode  dummy_outlier  \\\n",
       "0               0             0                  1              0   \n",
       "1               0             0                  1              0   \n",
       "2               0             1                  0              0   \n",
       "3               0             0                  1              0   \n",
       "4               0             0                  1              0   \n",
       "...           ...           ...                ...            ...   \n",
       "23086           0             0                  1              0   \n",
       "23087           0             0                  1              0   \n",
       "23088           0             0                  1              0   \n",
       "23089           0             0                  1              0   \n",
       "23090           0             1                  0              0   \n",
       "\n",
       "       dummy_very_long  \n",
       "0                    0  \n",
       "1                    0  \n",
       "2                    0  \n",
       "3                    0  \n",
       "4                    0  \n",
       "...                ...  \n",
       "23086                0  \n",
       "23087                0  \n",
       "23088                0  \n",
       "23089                0  \n",
       "23090                0  \n",
       "\n",
       "[23091 rows x 5 columns]"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Add the one-hot encoding to the dataset\n",
    "data_ICM_length = data_ICM_length.join(one_hot)\n",
    "# Remove the previous columns\n",
    "data_ICM_length.drop(columns=[\"data\",\"dummy\"], inplace=True)\n",
    "# Final dataset\n",
    "data_ICM_length"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "b0e5255d",
   "metadata": {},
   "source": [
    "#### Do the same for the ICM_type"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "id": "02f3f30b",
   "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>dummy_1</th>\n",
       "      <th>dummy_2</th>\n",
       "      <th>dummy_3</th>\n",
       "      <th>dummy_4</th>\n",
       "      <th>dummy_7</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23086</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23087</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23088</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23089</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23090</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>23091 rows × 5 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       dummy_1  dummy_2  dummy_3  dummy_4  dummy_7\n",
       "0            1        0        0        0        0\n",
       "1            0        0        1        0        0\n",
       "2            0        0        0        1        0\n",
       "3            1        0        0        0        0\n",
       "4            0        0        1        0        0\n",
       "...        ...      ...      ...      ...      ...\n",
       "23086        1        0        0        0        0\n",
       "23087        0        1        0        0        0\n",
       "23088        1        0        0        0        0\n",
       "23089        1        0        0        0        0\n",
       "23090        0        0        0        1        0\n",
       "\n",
       "[23091 rows x 5 columns]"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "one_hot = pd.get_dummies(data_ICM_type.feature_id, prefix='dummy')\n",
    "data_ICM_type = data_ICM_type.join(one_hot)\n",
    "data_ICM_type.drop(columns=[\"feature_id\"], inplace=True)\n",
    "data_ICM_type"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "71a6a1e0",
   "metadata": {},
   "source": [
    "#### Join all the encodings in a single dataset"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 43,
   "id": "c467b012",
   "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>dummy_1</th>\n",
       "      <th>dummy_2</th>\n",
       "      <th>dummy_3</th>\n",
       "      <th>dummy_4</th>\n",
       "      <th>dummy_7</th>\n",
       "      <th>dummy_long</th>\n",
       "      <th>dummy_medium</th>\n",
       "      <th>dummy_one_episode</th>\n",
       "      <th>dummy_outlier</th>\n",
       "      <th>dummy_very_long</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23086</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23087</th>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23088</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23089</th>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23090</th>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>23091 rows × 10 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "       dummy_1  dummy_2  dummy_3  dummy_4  dummy_7  dummy_long  dummy_medium  \\\n",
       "0            1        0        0        0        0           0             0   \n",
       "1            0        0        1        0        0           0             0   \n",
       "2            0        0        0        1        0           0             1   \n",
       "3            1        0        0        0        0           0             0   \n",
       "4            0        0        1        0        0           0             0   \n",
       "...        ...      ...      ...      ...      ...         ...           ...   \n",
       "23086        1        0        0        0        0           0             0   \n",
       "23087        0        1        0        0        0           0             0   \n",
       "23088        1        0        0        0        0           0             0   \n",
       "23089        1        0        0        0        0           0             0   \n",
       "23090        0        0        0        1        0           0             1   \n",
       "\n",
       "       dummy_one_episode  dummy_outlier  dummy_very_long  \n",
       "0                      1              0                0  \n",
       "1                      1              0                0  \n",
       "2                      0              0                0  \n",
       "3                      1              0                0  \n",
       "4                      1              0                0  \n",
       "...                  ...            ...              ...  \n",
       "23086                  1              0                0  \n",
       "23087                  1              0                0  \n",
       "23088                  1              0                0  \n",
       "23089                  1              0                0  \n",
       "23090                  0              0                0  \n",
       "\n",
       "[23091 rows x 10 columns]"
      ]
     },
     "execution_count": 43,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "ICM_complete = pd.DataFrame(data_ICM_type)\n",
    "ICM_complete = ICM_complete.join(data_ICM_length)\n",
    "ICM_complete"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 46,
   "id": "5f71a4d3",
   "metadata": {},
   "outputs": [],
   "source": [
    "# Save the dataset as a csv\n",
    "#ICM_complete.to_csv('data/ICM_one_hot.csv')  "
   ]
  }
 ],
 "metadata": {
  "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": "3.8.15"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}