ELI_data_frame_part1.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## NEW REPLACEMENT\n",
    "## Exploring ELI data frame\n",
    "#### Ting-Wei Shen, tis50@pitt.edu, Mar19, 2019\n",
    "\n",
    "### Table of contents\n",
    "- [1. Loading chosen data](#Loading-chosen-data) is where I combined files into one dataframe.\n",
    "- [2. Data Analysis](#Data-Analysis) is where I did some data calculation.\n",
    "- [3. Preprocessing for Linguistic Analysis](#Preprocessing-for-Linguistic-Analysis) is where I did some data calculation.\n",
    "- [4. Conclusion](#Conclusion) rounds up with some observations."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Loading chosen data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Pretty printing has been turned OFF\n"
     ]
    }
   ],
   "source": [
    "# Import all packages on top\n",
    "import nltk, glob, re\n",
    "import pandas as pd\n",
    "import matplotlib.pyplot as plt\n",
    "from scipy import stats\n",
    "import seaborn as sns\n",
    "\n",
    "# Turn off pretty printing \n",
    "%pprint"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [],
   "source": [
    "sns.set_style('darkgrid')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- First, check about answer.csv"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(46239, 6)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>answer_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>anon_id</th>\n",
       "      <th>user_file_id</th>\n",
       "      <th>text</th>\n",
       "      <th>is_deleted</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>5</td>\n",
       "      <td>eq0</td>\n",
       "      <td>7505</td>\n",
       "      <td>I met my friend Nife while I was studying in a...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>5</td>\n",
       "      <td>am8</td>\n",
       "      <td>7506</td>\n",
       "      <td>Ten years ago, I met a women on the train betw...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>12</td>\n",
       "      <td>dk5</td>\n",
       "      <td>7507</td>\n",
       "      <td>In my country we usually don't use tea bags. F...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>13</td>\n",
       "      <td>dk5</td>\n",
       "      <td>7507</td>\n",
       "      <td>I organized the instructions by time.</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>12</td>\n",
       "      <td>ad1</td>\n",
       "      <td>7508</td>\n",
       "      <td>First, prepare a port, loose tea, and cup.\\nSe...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>6</td>\n",
       "      <td>13</td>\n",
       "      <td>ad1</td>\n",
       "      <td>7508</td>\n",
       "      <td>By time</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>7</td>\n",
       "      <td>12</td>\n",
       "      <td>eg5</td>\n",
       "      <td>7509</td>\n",
       "      <td>First, prepare your cup, loose tea or bag tea,...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>8</td>\n",
       "      <td>13</td>\n",
       "      <td>eg5</td>\n",
       "      <td>7509</td>\n",
       "      <td>I organized the instructions by time, beacause...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>9</td>\n",
       "      <td>12</td>\n",
       "      <td>ad1</td>\n",
       "      <td>7510</td>\n",
       "      <td>First, prepare a port, loose tea, and cup.\\nSe...</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>10</td>\n",
       "      <td>13</td>\n",
       "      <td>ad1</td>\n",
       "      <td>7510</td>\n",
       "      <td>By time</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   answer_id  question_id anon_id  user_file_id  \\\n",
       "0          1            5     eq0          7505   \n",
       "1          2            5     am8          7506   \n",
       "2          3           12     dk5          7507   \n",
       "3          4           13     dk5          7507   \n",
       "4          5           12     ad1          7508   \n",
       "5          6           13     ad1          7508   \n",
       "6          7           12     eg5          7509   \n",
       "7          8           13     eg5          7509   \n",
       "8          9           12     ad1          7510   \n",
       "9         10           13     ad1          7510   \n",
       "\n",
       "                                                text  is_deleted  \n",
       "0  I met my friend Nife while I was studying in a...           0  \n",
       "1  Ten years ago, I met a women on the train betw...           0  \n",
       "2  In my country we usually don't use tea bags. F...           0  \n",
       "3              I organized the instructions by time.           0  \n",
       "4  First, prepare a port, loose tea, and cup.\\nSe...           0  \n",
       "5                                            By time           0  \n",
       "6  First, prepare your cup, loose tea or bag tea,...           0  \n",
       "7  I organized the instructions by time, beacause...           0  \n",
       "8  First, prepare a port, loose tea, and cup.\\nSe...           0  \n",
       "9                                            By time           0  "
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_answer = pd.read_csv(\"./data/answer.csv\")\n",
    "print(df_answer.shape)\n",
    "df_answer.head(10)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- I choose student_id(anon_id), question_id and text data to help me know the dataframe. Who write the essay with what kind of questions. \n",
    "\n",
    "- What we need are anon_id, question_id and text data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(46239, 3)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>anon_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>eq0</td>\n",
       "      <td>5</td>\n",
       "      <td>I met my friend Nife while I was studying in a...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>am8</td>\n",
       "      <td>5</td>\n",
       "      <td>Ten years ago, I met a women on the train betw...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>dk5</td>\n",
       "      <td>12</td>\n",
       "      <td>In my country we usually don't use tea bags. F...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>dk5</td>\n",
       "      <td>13</td>\n",
       "      <td>I organized the instructions by time.</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ad1</td>\n",
       "      <td>12</td>\n",
       "      <td>First, prepare a port, loose tea, and cup.\\nSe...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  anon_id  question_id                                               text\n",
       "0     eq0            5  I met my friend Nife while I was studying in a...\n",
       "1     am8            5  Ten years ago, I met a women on the train betw...\n",
       "2     dk5           12  In my country we usually don't use tea bags. F...\n",
       "3     dk5           13              I organized the instructions by time.\n",
       "4     ad1           12  First, prepare a port, loose tea, and cup.\\nSe..."
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli = pd.DataFrame()\n",
    "df_eli = df_answer[['anon_id', 'question_id', 'text']]\n",
    "print(df_eli.shape)\n",
    "df_eli.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Second, check about student information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(913, 21)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>anon_id</th>\n",
       "      <th>gender</th>\n",
       "      <th>birth_year</th>\n",
       "      <th>native_language</th>\n",
       "      <th>language_used_at_home</th>\n",
       "      <th>language_used_at_home_now</th>\n",
       "      <th>non_native_language_1</th>\n",
       "      <th>yrs_of_study_lang1</th>\n",
       "      <th>study_in_classroom_lang1</th>\n",
       "      <th>ways_of_study_lang1</th>\n",
       "      <th>...</th>\n",
       "      <th>yrs_of_study_lang2</th>\n",
       "      <th>study_in_classroom_lang2</th>\n",
       "      <th>ways_of_study_lang2</th>\n",
       "      <th>non_native_language_3</th>\n",
       "      <th>yrs_of_study_lang3</th>\n",
       "      <th>study_in_classroom_lang3</th>\n",
       "      <th>ways_of_study_lang3</th>\n",
       "      <th>createddate</th>\n",
       "      <th>modifieddate</th>\n",
       "      <th>course_history</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>ez9</td>\n",
       "      <td>Male</td>\n",
       "      <td>1978</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>NaN</td>\n",
       "      <td>English</td>\n",
       "      <td>more than 5 years</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Worked in pairs/groups;Studied...</td>\n",
       "      <td>...</td>\n",
       "      <td>less than 1 year</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Studied by myself</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2006-01-30 15:07:18</td>\n",
       "      <td>2006-03-14 15:13:37</td>\n",
       "      <td>6;12;18;24;30</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>gm3</td>\n",
       "      <td>Male</td>\n",
       "      <td>1980</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>NaN</td>\n",
       "      <td>English</td>\n",
       "      <td>more than 5 years</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Had a native-speaker teacher;S...</td>\n",
       "      <td>...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2006-01-30 15:07:28</td>\n",
       "      <td>2006-03-14 15:12:49</td>\n",
       "      <td>6;12;24;30;38</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>fg5</td>\n",
       "      <td>Male</td>\n",
       "      <td>1938</td>\n",
       "      <td>Nepali</td>\n",
       "      <td>Nepali</td>\n",
       "      <td>NaN</td>\n",
       "      <td>English</td>\n",
       "      <td>more than 5 years</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Worked in pairs/groups;Had a n...</td>\n",
       "      <td>...</td>\n",
       "      <td>less than 1 year</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Worked in pairs/groups;Had a n...</td>\n",
       "      <td>Hindi</td>\n",
       "      <td>more than 5 years</td>\n",
       "      <td>0.0</td>\n",
       "      <td>Studied by myself</td>\n",
       "      <td>2006-01-30 15:07:45</td>\n",
       "      <td>2006-03-14 15:11:36</td>\n",
       "      <td>18;24</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>ce5</td>\n",
       "      <td>Female</td>\n",
       "      <td>1984</td>\n",
       "      <td>Korean</td>\n",
       "      <td>Korean</td>\n",
       "      <td>NaN</td>\n",
       "      <td>English</td>\n",
       "      <td>more than 5 years</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Worked in pairs/groups;Had a n...</td>\n",
       "      <td>...</td>\n",
       "      <td>1-2 years</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Studied vocabulary;Listened to...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>0.0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>2006-01-30 15:07:49</td>\n",
       "      <td>2006-03-14 15:12:24</td>\n",
       "      <td>6;12;24;30;38;56</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>fi7</td>\n",
       "      <td>Female</td>\n",
       "      <td>1982</td>\n",
       "      <td>Korean</td>\n",
       "      <td>Korean;Japanese</td>\n",
       "      <td>NaN</td>\n",
       "      <td>English</td>\n",
       "      <td>more than 5 years</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Had a native-speaker teacher;S...</td>\n",
       "      <td>...</td>\n",
       "      <td>less than 1 year</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Studied vocabulary;Listened to...</td>\n",
       "      <td>French</td>\n",
       "      <td>1-2 years</td>\n",
       "      <td>1.0</td>\n",
       "      <td>Studied grammar;Studied vocabulary;Listened to...</td>\n",
       "      <td>2006-01-30 15:07:52</td>\n",
       "      <td>2006-03-14 15:12:17</td>\n",
       "      <td>6;12;24;30;38</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 21 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "  anon_id  gender  birth_year native_language language_used_at_home  \\\n",
       "0     ez9    Male        1978          Arabic                Arabic   \n",
       "1     gm3    Male        1980          Arabic                Arabic   \n",
       "2     fg5    Male        1938          Nepali                Nepali   \n",
       "3     ce5  Female        1984          Korean                Korean   \n",
       "4     fi7  Female        1982          Korean       Korean;Japanese   \n",
       "\n",
       "  language_used_at_home_now non_native_language_1 yrs_of_study_lang1  \\\n",
       "0                       NaN               English  more than 5 years   \n",
       "1                       NaN               English  more than 5 years   \n",
       "2                       NaN               English  more than 5 years   \n",
       "3                       NaN               English  more than 5 years   \n",
       "4                       NaN               English  more than 5 years   \n",
       "\n",
       "   study_in_classroom_lang1  \\\n",
       "0                       1.0   \n",
       "1                       1.0   \n",
       "2                       1.0   \n",
       "3                       1.0   \n",
       "4                       1.0   \n",
       "\n",
       "                                 ways_of_study_lang1        ...         \\\n",
       "0  Studied grammar;Worked in pairs/groups;Studied...        ...          \n",
       "1  Studied grammar;Had a native-speaker teacher;S...        ...          \n",
       "2  Studied grammar;Worked in pairs/groups;Had a n...        ...          \n",
       "3  Studied grammar;Worked in pairs/groups;Had a n...        ...          \n",
       "4  Studied grammar;Had a native-speaker teacher;S...        ...          \n",
       "\n",
       "  yrs_of_study_lang2 study_in_classroom_lang2  \\\n",
       "0   less than 1 year                      0.0   \n",
       "1                NaN                      0.0   \n",
       "2   less than 1 year                      1.0   \n",
       "3          1-2 years                      1.0   \n",
       "4   less than 1 year                      1.0   \n",
       "\n",
       "                                 ways_of_study_lang2 non_native_language_3  \\\n",
       "0                                  Studied by myself                   NaN   \n",
       "1                                                NaN                   NaN   \n",
       "2  Studied grammar;Worked in pairs/groups;Had a n...                 Hindi   \n",
       "3  Studied grammar;Studied vocabulary;Listened to...                   NaN   \n",
       "4  Studied grammar;Studied vocabulary;Listened to...                French   \n",
       "\n",
       "  yrs_of_study_lang3 study_in_classroom_lang3  \\\n",
       "0                NaN                      0.0   \n",
       "1                NaN                      0.0   \n",
       "2  more than 5 years                      0.0   \n",
       "3                NaN                      0.0   \n",
       "4          1-2 years                      1.0   \n",
       "\n",
       "                                 ways_of_study_lang3          createddate  \\\n",
       "0                                                NaN  2006-01-30 15:07:18   \n",
       "1                                                NaN  2006-01-30 15:07:28   \n",
       "2                                  Studied by myself  2006-01-30 15:07:45   \n",
       "3                                                NaN  2006-01-30 15:07:49   \n",
       "4  Studied grammar;Studied vocabulary;Listened to...  2006-01-30 15:07:52   \n",
       "\n",
       "          modifieddate    course_history  \n",
       "0  2006-03-14 15:13:37     6;12;18;24;30  \n",
       "1  2006-03-14 15:12:49     6;12;24;30;38  \n",
       "2  2006-03-14 15:11:36             18;24  \n",
       "3  2006-03-14 15:12:24  6;12;24;30;38;56  \n",
       "4  2006-03-14 15:12:17     6;12;24;30;38  \n",
       "\n",
       "[5 rows x 21 columns]"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_student = pd.read_csv(\"./data/student_information.csv\")\n",
    "print(df_student.shape)\n",
    "df_student.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Let's merge student information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(46239, 5)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>anon_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>eq0</td>\n",
       "      <td>5</td>\n",
       "      <td>I met my friend Nife while I was studying in a...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>am8</td>\n",
       "      <td>5</td>\n",
       "      <td>Ten years ago, I met a women on the train betw...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Thai</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>dk5</td>\n",
       "      <td>12</td>\n",
       "      <td>In my country we usually don't use tea bags. F...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Turkish</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>dk5</td>\n",
       "      <td>13</td>\n",
       "      <td>I organized the instructions by time.</td>\n",
       "      <td>Female</td>\n",
       "      <td>Turkish</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ad1</td>\n",
       "      <td>12</td>\n",
       "      <td>First, prepare a port, loose tea, and cup.\\nSe...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  anon_id  question_id                                               text  \\\n",
       "0     eq0            5  I met my friend Nife while I was studying in a...   \n",
       "1     am8            5  Ten years ago, I met a women on the train betw...   \n",
       "2     dk5           12  In my country we usually don't use tea bags. F...   \n",
       "3     dk5           13              I organized the instructions by time.   \n",
       "4     ad1           12  First, prepare a port, loose tea, and cup.\\nSe...   \n",
       "\n",
       "   gender native_language  \n",
       "0    Male          Arabic  \n",
       "1  Female            Thai  \n",
       "2  Female         Turkish  \n",
       "3  Female         Turkish  \n",
       "4  Female          Korean  "
      ]
     },
     "execution_count": 6,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_student_trim = df_student[['anon_id', 'gender', 'native_language']]\n",
    "df_eli = pd.merge(df_eli, df_student_trim, on='anon_id', how='left')\n",
    "print(df_eli.shape)\n",
    "df_eli.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "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>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>eq0</td>\n",
       "      <td>5</td>\n",
       "      <td>I met my friend Nife while I was studying in a...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>am8</td>\n",
       "      <td>5</td>\n",
       "      <td>Ten years ago, I met a women on the train betw...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Thai</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>dk5</td>\n",
       "      <td>12</td>\n",
       "      <td>In my country we usually don't use tea bags. F...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Turkish</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>dk5</td>\n",
       "      <td>13</td>\n",
       "      <td>I organized the instructions by time.</td>\n",
       "      <td>Female</td>\n",
       "      <td>Turkish</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ad1</td>\n",
       "      <td>12</td>\n",
       "      <td>First, prepare a port, loose tea, and cup.\\nSe...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  student_id  question_id                                               text  \\\n",
       "0        eq0            5  I met my friend Nife while I was studying in a...   \n",
       "1        am8            5  Ten years ago, I met a women on the train betw...   \n",
       "2        dk5           12  In my country we usually don't use tea bags. F...   \n",
       "3        dk5           13              I organized the instructions by time.   \n",
       "4        ad1           12  First, prepare a port, loose tea, and cup.\\nSe...   \n",
       "\n",
       "   gender native_language  \n",
       "0    Male          Arabic  \n",
       "1  Female            Thai  \n",
       "2  Female         Turkish  \n",
       "3  Female         Turkish  \n",
       "4  Female          Korean  "
      ]
     },
     "execution_count": 7,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Let's rename the column to help us recognize.\n",
    "df_eli = df_eli.rename(columns={'anon_id':'student_id'})\n",
    "df_eli.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- The dataframe looks great.\n",
    "\n",
    "- How about adding question column?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6138, 16)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>question_id</th>\n",
       "      <th>question_type_id</th>\n",
       "      <th>user_file_id</th>\n",
       "      <th>question_category_id</th>\n",
       "      <th>question_parent_id</th>\n",
       "      <th>explanation</th>\n",
       "      <th>stem</th>\n",
       "      <th>choiceA</th>\n",
       "      <th>choiceB</th>\n",
       "      <th>choiceC</th>\n",
       "      <th>choiceD</th>\n",
       "      <th>allow_text</th>\n",
       "      <th>order_num</th>\n",
       "      <th>is_deleted</th>\n",
       "      <th>org_level</th>\n",
       "      <th>direction</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>7430</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Answer the question.  Write a complete sentenc...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>2</td>\n",
       "      <td>7430</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Answer the question.  Write a complete sentenc...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>2</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>2</td>\n",
       "      <td>7430</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Answer the question.  Write a complete sentenc...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>3</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>4</td>\n",
       "      <td>7495</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Write an essay on ONE of the following topics....</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>1</td>\n",
       "      <td>7496</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Write a paragraph  about a relatioship that is...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>1</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   question_id  question_type_id  user_file_id  question_category_id  \\\n",
       "0            1                 2          7430                   NaN   \n",
       "1            2                 2          7430                   NaN   \n",
       "2            3                 2          7430                   NaN   \n",
       "3            4                 4          7495                   NaN   \n",
       "4            5                 1          7496                   NaN   \n",
       "\n",
       "   question_parent_id explanation  \\\n",
       "0                 NaN         NaN   \n",
       "1                 NaN         NaN   \n",
       "2                 NaN         NaN   \n",
       "3                 NaN         NaN   \n",
       "4                 NaN         NaN   \n",
       "\n",
       "                                                stem choiceA choiceB choiceC  \\\n",
       "0  Answer the question.  Write a complete sentenc...     NaN     NaN     NaN   \n",
       "1  Answer the question.  Write a complete sentenc...     NaN     NaN     NaN   \n",
       "2  Answer the question.  Write a complete sentenc...     NaN     NaN     NaN   \n",
       "3  Write an essay on ONE of the following topics....     NaN     NaN     NaN   \n",
       "4  Write a paragraph  about a relatioship that is...     NaN     NaN     NaN   \n",
       "\n",
       "  choiceD  allow_text  order_num  is_deleted  org_level direction  \n",
       "0     NaN           1          1           0        NaN       NaN  \n",
       "1     NaN           1          2           0        NaN       NaN  \n",
       "2     NaN           1          3           0        NaN       NaN  \n",
       "3     NaN           1          1           0        NaN       NaN  \n",
       "4     NaN           1          1           0        NaN       NaN  "
      ]
     },
     "execution_count": 8,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_question = pd.read_csv(\"./data/question.csv\")\n",
    "print(df_question.shape)\n",
    "df_question.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Select question_id and stem(question) only."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(6138, 2)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>question_id</th>\n",
       "      <th>question</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>1</td>\n",
       "      <td>Answer the question.  Write a complete sentenc...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>2</td>\n",
       "      <td>Answer the question.  Write a complete sentenc...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>3</td>\n",
       "      <td>Answer the question.  Write a complete sentenc...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>4</td>\n",
       "      <td>Write an essay on ONE of the following topics....</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>5</td>\n",
       "      <td>Write a paragraph  about a relatioship that is...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   question_id                                           question\n",
       "0            1  Answer the question.  Write a complete sentenc...\n",
       "1            2  Answer the question.  Write a complete sentenc...\n",
       "2            3  Answer the question.  Write a complete sentenc...\n",
       "3            4  Write an essay on ONE of the following topics....\n",
       "4            5  Write a paragraph  about a relatioship that is..."
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_question_trim = df_question[['question_id', 'stem']]\n",
    "df_question_trim = df_question_trim.rename(columns={'stem':'question'})\n",
    "print(df_question_trim.shape)\n",
    "df_question_trim.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(46239, 6)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>eq0</td>\n",
       "      <td>5</td>\n",
       "      <td>I met my friend Nife while I was studying in a...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Write a paragraph  about a relatioship that is...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>am8</td>\n",
       "      <td>5</td>\n",
       "      <td>Ten years ago, I met a women on the train betw...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Thai</td>\n",
       "      <td>Write a paragraph  about a relatioship that is...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>dk5</td>\n",
       "      <td>12</td>\n",
       "      <td>In my country we usually don't use tea bags. F...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Turkish</td>\n",
       "      <td>In five sentences or less, give instructions o...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>dk5</td>\n",
       "      <td>13</td>\n",
       "      <td>I organized the instructions by time.</td>\n",
       "      <td>Female</td>\n",
       "      <td>Turkish</td>\n",
       "      <td>How do you organize the instructions: by time,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>ad1</td>\n",
       "      <td>12</td>\n",
       "      <td>First, prepare a port, loose tea, and cup.\\nSe...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "      <td>In five sentences or less, give instructions o...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "  student_id  question_id                                               text  \\\n",
       "0        eq0            5  I met my friend Nife while I was studying in a...   \n",
       "1        am8            5  Ten years ago, I met a women on the train betw...   \n",
       "2        dk5           12  In my country we usually don't use tea bags. F...   \n",
       "3        dk5           13              I organized the instructions by time.   \n",
       "4        ad1           12  First, prepare a port, loose tea, and cup.\\nSe...   \n",
       "\n",
       "   gender native_language                                           question  \n",
       "0    Male          Arabic  Write a paragraph  about a relatioship that is...  \n",
       "1  Female            Thai  Write a paragraph  about a relatioship that is...  \n",
       "2  Female         Turkish  In five sentences or less, give instructions o...  \n",
       "3  Female         Turkish  How do you organize the instructions: by time,...  \n",
       "4  Female          Korean  In five sentences or less, give instructions o...  "
      ]
     },
     "execution_count": 10,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli = pd.merge(df_eli, df_question_trim, on='question_id', how='left')\n",
    "print(df_eli.shape)\n",
    "df_eli.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Since it is private data, I just choose sample data - about 500 to represent.\n",
    "- Then store the data as a new csv file."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_eli_short = df_eli.sample(n = 500, random_state = 1)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(500, 6)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>43439</th>\n",
       "      <td>gu2</td>\n",
       "      <td>5796</td>\n",
       "      <td>They see the online shopping do pollution but ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>753</th>\n",
       "      <td>cq4</td>\n",
       "      <td>37</td>\n",
       "      <td>1-Mismanagement\\nSentence: So if they were nom...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>For each of the following words, write the sen...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45003</th>\n",
       "      <td>ay8</td>\n",
       "      <td>5990</td>\n",
       "      <td>I was working out in the gym.GHABBAN.wav.mp3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What were  you doing at 8:00 this morning?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16101</th>\n",
       "      <td>ff4</td>\n",
       "      <td>2358</td>\n",
       "      <td>it's getting out of control. Something you sta...</td>\n",
       "      <td>Male</td>\n",
       "      <td>French</td>\n",
       "      <td>What does the phrase \"it's getting out of hand...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23749</th>\n",
       "      <td>cp5</td>\n",
       "      <td>3229</td>\n",
       "      <td>This tree has a very rigid trunk, I can not cu...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Write an original creative sentence using a vo...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      student_id  question_id  \\\n",
       "43439        gu2         5796   \n",
       "753          cq4           37   \n",
       "45003        ay8         5990   \n",
       "16101        ff4         2358   \n",
       "23749        cp5         3229   \n",
       "\n",
       "                                                    text  gender  \\\n",
       "43439  They see the online shopping do pollution but ...     NaN   \n",
       "753    1-Mismanagement\\nSentence: So if they were nom...  Female   \n",
       "45003       I was working out in the gym.GHABBAN.wav.mp3     NaN   \n",
       "16101  it's getting out of control. Something you sta...    Male   \n",
       "23749  This tree has a very rigid trunk, I can not cu...    Male   \n",
       "\n",
       "      native_language                                           question  \n",
       "43439             NaN  Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...  \n",
       "753            Arabic  For each of the following words, write the sen...  \n",
       "45003             NaN         What were  you doing at 8:00 this morning?  \n",
       "16101          French  What does the phrase \"it's getting out of hand...  \n",
       "23749          Arabic  Write an original creative sentence using a vo...  "
      ]
     },
     "execution_count": 12,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(df_eli_short.shape)\n",
    "df_eli_short.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- For the fair use, I have to eliminate personal information, so I just look through every text data. And mark the data with personal information."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "----------------------------\n",
      "0\n",
      "They see the online shopping do pollution but I am disagree with the author.\n",
      "I had heard that it is more energy efficient to do online shopping, even home delivery of groceries. Having products delivered to your home would have a smaller carbon footprint since. there is a single truck bringing goods to several people's homes as opposed to those individual people all getting in their cars and getting their items separately. I think similar to public transportation. When you have a communal way of transporting multiple things or people, it's better than each thing or person travelling separately.\n",
      "----------------------------\n",
      "1\n",
      "1-Mismanagement\n",
      "Sentence: So if they were nominated there would be no question of mismanagement of funds.\n",
      "Part of speech: Noun\n",
      "Definition: he is in charge financial. \n",
      "2-Demonstration\n",
      "Sentence: In fact, they claim that Victor was helping to organize a demonstration against the employment policies of Crowly Engineering. \n",
      "Part of speech: Noun\n",
      "Definition: An action, fact, etc .that proves that someone or something has particular quality, ability, emotion, etc.\n",
      "3-Orgainaize\n",
      "Sentence: E.X. 2\n",
      "Part of speech: Verb\n",
      "Definition: To form UNION or persuade people to join one.\n",
      "4-Disagreement\n",
      "Sentence: Mr. Crowly and his son, Victor had some very public disagreements on the construction of government housing for the poor.\n",
      "Part of speech: Noun\n",
      "Definition: A situation in which people express different opinions about something .\n",
      "5-Religious\n",
      "Sentence: Interviews with their friends show that they did not want their children to attend religious school.\n",
      "6-Construction\n",
      "Sentence: E.X. 4\n",
      "Part of speech: Noun\n",
      "Definition: The process or method of building something large such as a house, road.\n",
      "----------------------------\n",
      "2\n",
      "I was working out in the gym.GHABBAN.wav.mp3\n",
      "----------------------------\n",
      "3\n",
      "it's getting out of control. Something you start and then don't master it.\n",
      "----------------------------\n",
      "4\n",
      "This tree has a very rigid trunk, I can not cut by saw.\n"
     ]
    }
   ],
   "source": [
    "# print all data with index count, mark the data with personal information\n",
    "# just print first 5 to represent\n",
    "# mark index 112,117,251,352,354,392,424,441\n",
    "count = 0\n",
    "for x in df_eli_short.text[:5]:\n",
    "    print('----------------------------')\n",
    "    print(count)\n",
    "    print(x)\n",
    "    count = count + 1"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Drop personal information"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_eli_short = df_eli_short.drop(df_eli_short.index[[112,117,251,352,354,392,424,441]])"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- It looks like the dataframe is good"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(492, 6)\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>43439</th>\n",
       "      <td>gu2</td>\n",
       "      <td>5796</td>\n",
       "      <td>They see the online shopping do pollution but ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>753</th>\n",
       "      <td>cq4</td>\n",
       "      <td>37</td>\n",
       "      <td>1-Mismanagement\\nSentence: So if they were nom...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>For each of the following words, write the sen...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45003</th>\n",
       "      <td>ay8</td>\n",
       "      <td>5990</td>\n",
       "      <td>I was working out in the gym.GHABBAN.wav.mp3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What were  you doing at 8:00 this morning?</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16101</th>\n",
       "      <td>ff4</td>\n",
       "      <td>2358</td>\n",
       "      <td>it's getting out of control. Something you sta...</td>\n",
       "      <td>Male</td>\n",
       "      <td>French</td>\n",
       "      <td>What does the phrase \"it's getting out of hand...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23749</th>\n",
       "      <td>cp5</td>\n",
       "      <td>3229</td>\n",
       "      <td>This tree has a very rigid trunk, I can not cu...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Write an original creative sentence using a vo...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      student_id  question_id  \\\n",
       "43439        gu2         5796   \n",
       "753          cq4           37   \n",
       "45003        ay8         5990   \n",
       "16101        ff4         2358   \n",
       "23749        cp5         3229   \n",
       "\n",
       "                                                    text  gender  \\\n",
       "43439  They see the online shopping do pollution but ...     NaN   \n",
       "753    1-Mismanagement\\nSentence: So if they were nom...  Female   \n",
       "45003       I was working out in the gym.GHABBAN.wav.mp3     NaN   \n",
       "16101  it's getting out of control. Something you sta...    Male   \n",
       "23749  This tree has a very rigid trunk, I can not cu...    Male   \n",
       "\n",
       "      native_language                                           question  \n",
       "43439             NaN  Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...  \n",
       "753            Arabic  For each of the following words, write the sen...  \n",
       "45003             NaN         What were  you doing at 8:00 this morning?  \n",
       "16101          French  What does the phrase \"it's getting out of hand...  \n",
       "23749          Arabic  Write an original creative sentence using a vo...  "
      ]
     },
     "execution_count": 15,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "print(df_eli_short.shape)\n",
    "df_eli_short.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- save the file as sample data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [],
   "source": [
    "#df_eli_short.to_csv(\"./data/df_eli_short.csv\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Data Analysis\n",
    "### 1. Text length as token count\n",
    "- Let's count text length\n",
    "- In this part, nltk.word_tokenize return error caused by the wrong input first. That is because some text is stored in \"float\" type, and word_tokenize only accepts \"string\" format. I have to set every text into \"string\" format first, then utilize word_tokenize()."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [],
   "source": [
    "def getTokens(txt):\n",
    "    txt = str(txt)\n",
    "    x = nltk.word_tokenize(txt)\n",
    "    return x"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_eli_short['tokens'] = df_eli_short.text.map(getTokens)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 19,
   "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>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "      <th>tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>43439</th>\n",
       "      <td>gu2</td>\n",
       "      <td>5796</td>\n",
       "      <td>They see the online shopping do pollution but ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...</td>\n",
       "      <td>[They, see, the, online, shopping, do, polluti...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>753</th>\n",
       "      <td>cq4</td>\n",
       "      <td>37</td>\n",
       "      <td>1-Mismanagement\\nSentence: So if they were nom...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>For each of the following words, write the sen...</td>\n",
       "      <td>[1-Mismanagement, Sentence, :, So, if, they, w...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45003</th>\n",
       "      <td>ay8</td>\n",
       "      <td>5990</td>\n",
       "      <td>I was working out in the gym.GHABBAN.wav.mp3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What were  you doing at 8:00 this morning?</td>\n",
       "      <td>[I, was, working, out, in, the, gym.GHABBAN.wa...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16101</th>\n",
       "      <td>ff4</td>\n",
       "      <td>2358</td>\n",
       "      <td>it's getting out of control. Something you sta...</td>\n",
       "      <td>Male</td>\n",
       "      <td>French</td>\n",
       "      <td>What does the phrase \"it's getting out of hand...</td>\n",
       "      <td>[it, 's, getting, out, of, control, ., Somethi...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23749</th>\n",
       "      <td>cp5</td>\n",
       "      <td>3229</td>\n",
       "      <td>This tree has a very rigid trunk, I can not cu...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Write an original creative sentence using a vo...</td>\n",
       "      <td>[This, tree, has, a, very, rigid, trunk, ,, I,...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      student_id  question_id  \\\n",
       "43439        gu2         5796   \n",
       "753          cq4           37   \n",
       "45003        ay8         5990   \n",
       "16101        ff4         2358   \n",
       "23749        cp5         3229   \n",
       "\n",
       "                                                    text  gender  \\\n",
       "43439  They see the online shopping do pollution but ...     NaN   \n",
       "753    1-Mismanagement\\nSentence: So if they were nom...  Female   \n",
       "45003       I was working out in the gym.GHABBAN.wav.mp3     NaN   \n",
       "16101  it's getting out of control. Something you sta...    Male   \n",
       "23749  This tree has a very rigid trunk, I can not cu...    Male   \n",
       "\n",
       "      native_language                                           question  \\\n",
       "43439             NaN  Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...   \n",
       "753            Arabic  For each of the following words, write the sen...   \n",
       "45003             NaN         What were  you doing at 8:00 this morning?   \n",
       "16101          French  What does the phrase \"it's getting out of hand...   \n",
       "23749          Arabic  Write an original creative sentence using a vo...   \n",
       "\n",
       "                                                  tokens  \n",
       "43439  [They, see, the, online, shopping, do, polluti...  \n",
       "753    [1-Mismanagement, Sentence, :, So, if, they, w...  \n",
       "45003  [I, was, working, out, in, the, gym.GHABBAN.wa...  \n",
       "16101  [it, 's, getting, out, of, control, ., Somethi...  \n",
       "23749  [This, tree, has, a, very, rigid, trunk, ,, I,...  "
      ]
     },
     "execution_count": 19,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 20,
   "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>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "      <th>tokens</th>\n",
       "      <th>token_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>43439</th>\n",
       "      <td>gu2</td>\n",
       "      <td>5796</td>\n",
       "      <td>They see the online shopping do pollution but ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...</td>\n",
       "      <td>[They, see, the, online, shopping, do, polluti...</td>\n",
       "      <td>109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>753</th>\n",
       "      <td>cq4</td>\n",
       "      <td>37</td>\n",
       "      <td>1-Mismanagement\\nSentence: So if they were nom...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>For each of the following words, write the sen...</td>\n",
       "      <td>[1-Mismanagement, Sentence, :, So, if, they, w...</td>\n",
       "      <td>198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45003</th>\n",
       "      <td>ay8</td>\n",
       "      <td>5990</td>\n",
       "      <td>I was working out in the gym.GHABBAN.wav.mp3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What were  you doing at 8:00 this morning?</td>\n",
       "      <td>[I, was, working, out, in, the, gym.GHABBAN.wa...</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16101</th>\n",
       "      <td>ff4</td>\n",
       "      <td>2358</td>\n",
       "      <td>it's getting out of control. Something you sta...</td>\n",
       "      <td>Male</td>\n",
       "      <td>French</td>\n",
       "      <td>What does the phrase \"it's getting out of hand...</td>\n",
       "      <td>[it, 's, getting, out, of, control, ., Somethi...</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23749</th>\n",
       "      <td>cp5</td>\n",
       "      <td>3229</td>\n",
       "      <td>This tree has a very rigid trunk, I can not cu...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Write an original creative sentence using a vo...</td>\n",
       "      <td>[This, tree, has, a, very, rigid, trunk, ,, I,...</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      student_id  question_id  \\\n",
       "43439        gu2         5796   \n",
       "753          cq4           37   \n",
       "45003        ay8         5990   \n",
       "16101        ff4         2358   \n",
       "23749        cp5         3229   \n",
       "\n",
       "                                                    text  gender  \\\n",
       "43439  They see the online shopping do pollution but ...     NaN   \n",
       "753    1-Mismanagement\\nSentence: So if they were nom...  Female   \n",
       "45003       I was working out in the gym.GHABBAN.wav.mp3     NaN   \n",
       "16101  it's getting out of control. Something you sta...    Male   \n",
       "23749  This tree has a very rigid trunk, I can not cu...    Male   \n",
       "\n",
       "      native_language                                           question  \\\n",
       "43439             NaN  Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...   \n",
       "753            Arabic  For each of the following words, write the sen...   \n",
       "45003             NaN         What were  you doing at 8:00 this morning?   \n",
       "16101          French  What does the phrase \"it's getting out of hand...   \n",
       "23749          Arabic  Write an original creative sentence using a vo...   \n",
       "\n",
       "                                                  tokens  token_count  \n",
       "43439  [They, see, the, online, shopping, do, polluti...          109  \n",
       "753    [1-Mismanagement, Sentence, :, So, if, they, w...          198  \n",
       "45003  [I, was, working, out, in, the, gym.GHABBAN.wa...            7  \n",
       "16101  [it, 's, getting, out, of, control, ., Somethi...           17  \n",
       "23749  [This, tree, has, a, very, rigid, trunk, ,, I,...           15  "
      ]
     },
     "execution_count": 20,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short['token_count'] = df_eli_short.tokens.map(len)\n",
    "df_eli_short.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- There are a lot of short text that less than 100 words. If we took a look at the questions, we can understand that students may be asked to answer just in one word.\n",
    "- For example, question_id '1866' - 'Look at the word \"warriors\" in paragraph 14.  Does it mean a) soldiers or b) hard workers?'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "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>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "      <th>tokens</th>\n",
       "      <th>token_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>12368</th>\n",
       "      <td>dz9</td>\n",
       "      <td>1866</td>\n",
       "      <td>a</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>Look at the word \"warriors\" in paragraph 14.  ...</td>\n",
       "      <td>[a]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10979</th>\n",
       "      <td>ec9</td>\n",
       "      <td>1651</td>\n",
       "      <td>a</td>\n",
       "      <td>Male</td>\n",
       "      <td>Korean</td>\n",
       "      <td>Look at the word \"compounded\" in paragraph 13....</td>\n",
       "      <td>[a]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19519</th>\n",
       "      <td>cn1</td>\n",
       "      <td>2753</td>\n",
       "      <td>t</td>\n",
       "      <td>Female</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>T or F</td>\n",
       "      <td>[t]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27560</th>\n",
       "      <td>az3</td>\n",
       "      <td>3798</td>\n",
       "      <td>offended</td>\n",
       "      <td>Female</td>\n",
       "      <td>Taiwanese</td>\n",
       "      <td>Choose a vocabulary word from the article and ...</td>\n",
       "      <td>[offended]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8498</th>\n",
       "      <td>bm1</td>\n",
       "      <td>1304</td>\n",
       "      <td>a</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>American Proverb: “The way to ______ man’s hea...</td>\n",
       "      <td>[a]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>38423</th>\n",
       "      <td>fm5</td>\n",
       "      <td>5123</td>\n",
       "      <td>supplanted</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>dislocated</td>\n",
       "      <td>[supplanted]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35611</th>\n",
       "      <td>ao2</td>\n",
       "      <td>4816</td>\n",
       "      <td>C</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>The topic of the third reading is\\n\\na.  the s...</td>\n",
       "      <td>[C]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10531</th>\n",
       "      <td>aa8</td>\n",
       "      <td>1630</td>\n",
       "      <td>c</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "      <td>to test out (6)\\n\\na.  to stop something\\nb.  ...</td>\n",
       "      <td>[c]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18994</th>\n",
       "      <td>cy6</td>\n",
       "      <td>2679</td>\n",
       "      <td>is</td>\n",
       "      <td>Female</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>Grammar _____ (be) the best class all day!</td>\n",
       "      <td>[is]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>36348</th>\n",
       "      <td>ex3</td>\n",
       "      <td>4953</td>\n",
       "      <td>C</td>\n",
       "      <td>Male</td>\n",
       "      <td>Korean</td>\n",
       "      <td>It _______ right now.</td>\n",
       "      <td>[C]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16293</th>\n",
       "      <td>cx0</td>\n",
       "      <td>2401</td>\n",
       "      <td>judge</td>\n",
       "      <td>Female</td>\n",
       "      <td>Chinese</td>\n",
       "      <td>evaluate (v)</td>\n",
       "      <td>[judge]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26089</th>\n",
       "      <td>cm8</td>\n",
       "      <td>3566</td>\n",
       "      <td>Unless</td>\n",
       "      <td>Female</td>\n",
       "      <td>Russian</td>\n",
       "      <td>________ it stops raining, we won't be able to...</td>\n",
       "      <td>[Unless]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5094</th>\n",
       "      <td>gg2</td>\n",
       "      <td>767</td>\n",
       "      <td>c</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "      <td>When I was a child, my family ______ a horse. ...</td>\n",
       "      <td>[c]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5556</th>\n",
       "      <td>di3</td>\n",
       "      <td>737</td>\n",
       "      <td>finished</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Some students will begin university studies af...</td>\n",
       "      <td>[finished]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3925</th>\n",
       "      <td>bs1</td>\n",
       "      <td>525</td>\n",
       "      <td>that</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>My sister, ___ I haven't seen for 3 years, is ...</td>\n",
       "      <td>[that]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14332</th>\n",
       "      <td>aa8</td>\n",
       "      <td>2128</td>\n",
       "      <td>damage</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "      <td>feat, damage, accomplishment</td>\n",
       "      <td>[damage]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5236</th>\n",
       "      <td>fi1</td>\n",
       "      <td>775</td>\n",
       "      <td>a</td>\n",
       "      <td>Female</td>\n",
       "      <td>Chinese</td>\n",
       "      <td>Have you ever had _____ dream about winning th...</td>\n",
       "      <td>[a]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>35300</th>\n",
       "      <td>aw6</td>\n",
       "      <td>4811</td>\n",
       "      <td>c</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What does \"that\" in the last sentence of the s...</td>\n",
       "      <td>[c]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8850</th>\n",
       "      <td>dz9</td>\n",
       "      <td>1388</td>\n",
       "      <td>avoid</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>Luis becomes very sick when he eats corn.  He ...</td>\n",
       "      <td>[avoid]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9090</th>\n",
       "      <td>bm3</td>\n",
       "      <td>1450</td>\n",
       "      <td>c</td>\n",
       "      <td>Male</td>\n",
       "      <td>Korean</td>\n",
       "      <td>Look at the word \"argue\"  in paragraph 1, sent...</td>\n",
       "      <td>[c]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>44289</th>\n",
       "      <td>cq8</td>\n",
       "      <td>5918</td>\n",
       "      <td>5</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>How many ELI classes are you taking?</td>\n",
       "      <td>[5]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8409</th>\n",
       "      <td>fg9</td>\n",
       "      <td>1305</td>\n",
       "      <td>a</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>My dog is afraid of _______ lightning—she alwa...</td>\n",
       "      <td>[a]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>42990</th>\n",
       "      <td>db2</td>\n",
       "      <td>5728</td>\n",
       "      <td>initial</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>first, at the outset (i)</td>\n",
       "      <td>[initial]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8811</th>\n",
       "      <td>ff2</td>\n",
       "      <td>1393</td>\n",
       "      <td>courage</td>\n",
       "      <td>Female</td>\n",
       "      <td>Chinese</td>\n",
       "      <td>It is important for ELI students to have _____...</td>\n",
       "      <td>[courage]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13765</th>\n",
       "      <td>bm7</td>\n",
       "      <td>2058</td>\n",
       "      <td>Exciting</td>\n",
       "      <td>Female</td>\n",
       "      <td>Portuguese</td>\n",
       "      <td>What does the word \"vibrant\" (paragraphs 1 and...</td>\n",
       "      <td>[Exciting]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4143</th>\n",
       "      <td>ca5</td>\n",
       "      <td>511</td>\n",
       "      <td>Internet</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>What modern invention helps people take course...</td>\n",
       "      <td>[Internet]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3991</th>\n",
       "      <td>fi5</td>\n",
       "      <td>531</td>\n",
       "      <td>which</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "      <td>The book ___ your brother wants is on the table.</td>\n",
       "      <td>[which]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19610</th>\n",
       "      <td>db3</td>\n",
       "      <td>2753</td>\n",
       "      <td>T</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>T or F</td>\n",
       "      <td>[T]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23222</th>\n",
       "      <td>bu1</td>\n",
       "      <td>3173</td>\n",
       "      <td>actively</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>Students learn faster when they are _____ usin...</td>\n",
       "      <td>[actively]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8373</th>\n",
       "      <td>bs1</td>\n",
       "      <td>1305</td>\n",
       "      <td>0</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>My dog is afraid of _______ lightning—she alwa...</td>\n",
       "      <td>[0]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8253</th>\n",
       "      <td>fh9</td>\n",
       "      <td>1265</td>\n",
       "      <td>The</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>______ Middle East is the home of many ancient...</td>\n",
       "      <td>[The]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5199</th>\n",
       "      <td>bs1</td>\n",
       "      <td>772</td>\n",
       "      <td>b</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>There are only ___________ days left in the Sp...</td>\n",
       "      <td>[b]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8863</th>\n",
       "      <td>gv1</td>\n",
       "      <td>1390</td>\n",
       "      <td>inflexible</td>\n",
       "      <td>Male</td>\n",
       "      <td>Korean</td>\n",
       "      <td>The crime victim gave the police _____________...</td>\n",
       "      <td>[inflexible]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19041</th>\n",
       "      <td>gb1</td>\n",
       "      <td>2703</td>\n",
       "      <td>ate</td>\n",
       "      <td>Male</td>\n",
       "      <td>Other</td>\n",
       "      <td>I ___________________ (eat) ice cream for brea...</td>\n",
       "      <td>[ate]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13460</th>\n",
       "      <td>dc6</td>\n",
       "      <td>1991</td>\n",
       "      <td>b</td>\n",
       "      <td>Female</td>\n",
       "      <td>Spanish</td>\n",
       "      <td>What does the word \"opens\" mean in the first p...</td>\n",
       "      <td>[b]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3813</th>\n",
       "      <td>df0</td>\n",
       "      <td>468</td>\n",
       "      <td>where</td>\n",
       "      <td>Female</td>\n",
       "      <td>Japanese</td>\n",
       "      <td>You should go to Washington, DC, __________ th...</td>\n",
       "      <td>[where]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3907</th>\n",
       "      <td>ei2</td>\n",
       "      <td>527</td>\n",
       "      <td>who</td>\n",
       "      <td>Female</td>\n",
       "      <td>Chinese</td>\n",
       "      <td>I have 3 brothers.  The brother ___ lives in N...</td>\n",
       "      <td>[who]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5031</th>\n",
       "      <td>es9</td>\n",
       "      <td>770</td>\n",
       "      <td>d</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "      <td>______ the U.S. for the first time, the touris...</td>\n",
       "      <td>[d]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19508</th>\n",
       "      <td>cl1</td>\n",
       "      <td>2755</td>\n",
       "      <td>false</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>T or F\\nExplain:  What made you choose your an...</td>\n",
       "      <td>[false]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>41889</th>\n",
       "      <td>gx1</td>\n",
       "      <td>5591</td>\n",
       "      <td>10/23/2011</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Date the reading was written:</td>\n",
       "      <td>[10/23/2011]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28307</th>\n",
       "      <td>cp5</td>\n",
       "      <td>3905</td>\n",
       "      <td>Robbery</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Choose a vocabulary word from the article and ...</td>\n",
       "      <td>[Robbery]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4008</th>\n",
       "      <td>ey9</td>\n",
       "      <td>528</td>\n",
       "      <td>who</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>I want to practice my Spanish, but I don't kno...</td>\n",
       "      <td>[who]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5795</th>\n",
       "      <td>dd9</td>\n",
       "      <td>992</td>\n",
       "      <td>must</td>\n",
       "      <td>Male</td>\n",
       "      <td>Korean</td>\n",
       "      <td>The longest river in the world is \\na. the Nil...</td>\n",
       "      <td>[must]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5235</th>\n",
       "      <td>fi1</td>\n",
       "      <td>774</td>\n",
       "      <td>c</td>\n",
       "      <td>Female</td>\n",
       "      <td>Chinese</td>\n",
       "      <td>... 3 people ______ in a hit-and-run accident....</td>\n",
       "      <td>[c]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6888</th>\n",
       "      <td>ft2</td>\n",
       "      <td>1134</td>\n",
       "      <td>d</td>\n",
       "      <td>Female</td>\n",
       "      <td>Korean</td>\n",
       "      <td>What does \"reflect\" mean in paragraph 5? (Writ...</td>\n",
       "      <td>[d]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12112</th>\n",
       "      <td>bz1</td>\n",
       "      <td>1839</td>\n",
       "      <td>D</td>\n",
       "      <td>Male</td>\n",
       "      <td>Chinese</td>\n",
       "      <td>Why didn't the school put an air conditioner i...</td>\n",
       "      <td>[D]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19159</th>\n",
       "      <td>ef1</td>\n",
       "      <td>2733</td>\n",
       "      <td>yhryr</td>\n",
       "      <td>Male</td>\n",
       "      <td>Korean</td>\n",
       "      <td>Complete the Review Tests in you book on pages...</td>\n",
       "      <td>[yhryr]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19473</th>\n",
       "      <td>gd7</td>\n",
       "      <td>2759</td>\n",
       "      <td>T</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>T or F</td>\n",
       "      <td>[T]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3894</th>\n",
       "      <td>gk3</td>\n",
       "      <td>524</td>\n",
       "      <td>whose</td>\n",
       "      <td>Male</td>\n",
       "      <td>Turkish</td>\n",
       "      <td>I could see that Lisa was a wonderful woman, _...</td>\n",
       "      <td>[whose]</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      student_id  question_id         text  gender native_language  \\\n",
       "12368        dz9         1866            a  Female        Japanese   \n",
       "10979        ec9         1651            a    Male          Korean   \n",
       "19519        cn1         2753            t  Female         Spanish   \n",
       "27560        az3         3798     offended  Female       Taiwanese   \n",
       "8498         bm1         1304            a     NaN             NaN   \n",
       "38423        fm5         5123  supplanted     Male          Arabic   \n",
       "35611        ao2         4816            C    Male          Arabic   \n",
       "10531        aa8         1630            c  Female          Korean   \n",
       "18994        cy6         2679           is  Female         Spanish   \n",
       "36348        ex3         4953            C    Male          Korean   \n",
       "16293        cx0         2401        judge  Female         Chinese   \n",
       "26089        cm8         3566       Unless  Female         Russian   \n",
       "5094         gg2          767            c  Female          Korean   \n",
       "5556         di3          737     finished    Male          Arabic   \n",
       "3925         bs1          525         that  Female        Japanese   \n",
       "14332        aa8         2128       damage  Female          Korean   \n",
       "5236         fi1          775            a  Female         Chinese   \n",
       "35300        aw6         4811            c     NaN             NaN   \n",
       "8850         dz9         1388        avoid  Female        Japanese   \n",
       "9090         bm3         1450            c    Male          Korean   \n",
       "44289        cq8         5918            5     NaN             NaN   \n",
       "8409         fg9         1305            a    Male          Arabic   \n",
       "42990        db2         5728      initial     NaN             NaN   \n",
       "8811         ff2         1393      courage  Female         Chinese   \n",
       "13765        bm7         2058     Exciting  Female      Portuguese   \n",
       "4143         ca5          511     Internet    Male          Arabic   \n",
       "3991         fi5          531        which  Female          Korean   \n",
       "19610        db3         2753            T    Male          Arabic   \n",
       "23222        bu1         3173     actively  Female        Japanese   \n",
       "8373         bs1         1305            0  Female        Japanese   \n",
       "8253         fh9         1265          The  Female        Japanese   \n",
       "5199         bs1          772            b  Female        Japanese   \n",
       "8863         gv1         1390   inflexible    Male          Korean   \n",
       "19041        gb1         2703          ate    Male           Other   \n",
       "13460        dc6         1991            b  Female         Spanish   \n",
       "3813         df0          468        where  Female        Japanese   \n",
       "3907         ei2          527          who  Female         Chinese   \n",
       "5031         es9          770            d  Female          Korean   \n",
       "19508        cl1         2755        false    Male          Arabic   \n",
       "41889        gx1         5591   10/23/2011     NaN             NaN   \n",
       "28307        cp5         3905      Robbery    Male          Arabic   \n",
       "4008         ey9          528          who     NaN             NaN   \n",
       "5795         dd9          992         must    Male          Korean   \n",
       "5235         fi1          774            c  Female         Chinese   \n",
       "6888         ft2         1134            d  Female          Korean   \n",
       "12112        bz1         1839            D    Male         Chinese   \n",
       "19159        ef1         2733        yhryr    Male          Korean   \n",
       "19473        gd7         2759            T    Male          Arabic   \n",
       "3894         gk3          524        whose    Male         Turkish   \n",
       "\n",
       "                                                question        tokens  \\\n",
       "12368  Look at the word \"warriors\" in paragraph 14.  ...           [a]   \n",
       "10979  Look at the word \"compounded\" in paragraph 13....           [a]   \n",
       "19519                                             T or F           [t]   \n",
       "27560  Choose a vocabulary word from the article and ...    [offended]   \n",
       "8498   American Proverb: “The way to ______ man’s hea...           [a]   \n",
       "38423                                         dislocated  [supplanted]   \n",
       "35611  The topic of the third reading is\\n\\na.  the s...           [C]   \n",
       "10531  to test out (6)\\n\\na.  to stop something\\nb.  ...           [c]   \n",
       "18994         Grammar _____ (be) the best class all day!          [is]   \n",
       "36348                              It _______ right now.           [C]   \n",
       "16293                                       evaluate (v)       [judge]   \n",
       "26089  ________ it stops raining, we won't be able to...      [Unless]   \n",
       "5094   When I was a child, my family ______ a horse. ...           [c]   \n",
       "5556   Some students will begin university studies af...    [finished]   \n",
       "3925   My sister, ___ I haven't seen for 3 years, is ...        [that]   \n",
       "14332                       feat, damage, accomplishment      [damage]   \n",
       "5236   Have you ever had _____ dream about winning th...           [a]   \n",
       "35300  What does \"that\" in the last sentence of the s...           [c]   \n",
       "8850   Luis becomes very sick when he eats corn.  He ...       [avoid]   \n",
       "9090   Look at the word \"argue\"  in paragraph 1, sent...           [c]   \n",
       "44289               How many ELI classes are you taking?           [5]   \n",
       "8409   My dog is afraid of _______ lightning—she alwa...           [a]   \n",
       "42990                           first, at the outset (i)     [initial]   \n",
       "8811   It is important for ELI students to have _____...     [courage]   \n",
       "13765  What does the word \"vibrant\" (paragraphs 1 and...    [Exciting]   \n",
       "4143   What modern invention helps people take course...    [Internet]   \n",
       "3991    The book ___ your brother wants is on the table.       [which]   \n",
       "19610                                             T or F           [T]   \n",
       "23222  Students learn faster when they are _____ usin...    [actively]   \n",
       "8373   My dog is afraid of _______ lightning—she alwa...           [0]   \n",
       "8253   ______ Middle East is the home of many ancient...         [The]   \n",
       "5199   There are only ___________ days left in the Sp...           [b]   \n",
       "8863   The crime victim gave the police _____________...  [inflexible]   \n",
       "19041  I ___________________ (eat) ice cream for brea...         [ate]   \n",
       "13460  What does the word \"opens\" mean in the first p...           [b]   \n",
       "3813   You should go to Washington, DC, __________ th...       [where]   \n",
       "3907   I have 3 brothers.  The brother ___ lives in N...         [who]   \n",
       "5031   ______ the U.S. for the first time, the touris...           [d]   \n",
       "19508  T or F\\nExplain:  What made you choose your an...       [false]   \n",
       "41889                      Date the reading was written:  [10/23/2011]   \n",
       "28307  Choose a vocabulary word from the article and ...     [Robbery]   \n",
       "4008   I want to practice my Spanish, but I don't kno...         [who]   \n",
       "5795   The longest river in the world is \\na. the Nil...        [must]   \n",
       "5235   ... 3 people ______ in a hit-and-run accident....           [c]   \n",
       "6888   What does \"reflect\" mean in paragraph 5? (Writ...           [d]   \n",
       "12112  Why didn't the school put an air conditioner i...           [D]   \n",
       "19159  Complete the Review Tests in you book on pages...       [yhryr]   \n",
       "19473                                             T or F           [T]   \n",
       "3894   I could see that Lisa was a wonderful woman, _...       [whose]   \n",
       "\n",
       "       token_count  \n",
       "12368            1  \n",
       "10979            1  \n",
       "19519            1  \n",
       "27560            1  \n",
       "8498             1  \n",
       "38423            1  \n",
       "35611            1  \n",
       "10531            1  \n",
       "18994            1  \n",
       "36348            1  \n",
       "16293            1  \n",
       "26089            1  \n",
       "5094             1  \n",
       "5556             1  \n",
       "3925             1  \n",
       "14332            1  \n",
       "5236             1  \n",
       "35300            1  \n",
       "8850             1  \n",
       "9090             1  \n",
       "44289            1  \n",
       "8409             1  \n",
       "42990            1  \n",
       "8811             1  \n",
       "13765            1  \n",
       "4143             1  \n",
       "3991             1  \n",
       "19610            1  \n",
       "23222            1  \n",
       "8373             1  \n",
       "8253             1  \n",
       "5199             1  \n",
       "8863             1  \n",
       "19041            1  \n",
       "13460            1  \n",
       "3813             1  \n",
       "3907             1  \n",
       "5031             1  \n",
       "19508            1  \n",
       "41889            1  \n",
       "28307            1  \n",
       "4008             1  \n",
       "5795             1  \n",
       "5235             1  \n",
       "6888             1  \n",
       "12112            1  \n",
       "19159            1  \n",
       "19473            1  \n",
       "3894             1  "
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short[df_eli_short.token_count == 1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "49"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "len(df_eli_short[df_eli_short.token_count == 1])"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Look at the word \"warriors\" in paragraph 14.  Does it mean a) soldiers or b) hard workers?'"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short[df_eli_short.token_count == 1].iloc[0,5]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Student just answer 'a'. That is really reasonable.\n",
    "- However, this situation may cause that the whole dataframe contains many one word texts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 24,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'a'"
      ]
     },
     "execution_count": 24,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short[df_eli_short.token_count == 1].iloc[0,2]"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- If we counted the value by number of tokens, we can find that data has a lot of short texts."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 25,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1     49\n",
       "8     13\n",
       "5     13\n",
       "2     12\n",
       "6     12\n",
       "7     12\n",
       "15    11\n",
       "12    11\n",
       "16    11\n",
       "13    10\n",
       "Name: token_count, dtype: int64"
      ]
     },
     "execution_count": 25,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.token_count.value_counts().head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 26,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.262, 64.05]      315\n",
       "(64.05, 127.1]        56\n",
       "(127.1, 190.15]       47\n",
       "(190.15, 253.2]       19\n",
       "(253.2, 316.25]       13\n",
       "(442.35, 505.4]        7\n",
       "(316.25, 379.3]        7\n",
       "(379.3, 442.35]        7\n",
       "(568.45, 631.5]        5\n",
       "(631.5, 694.55]        4\n",
       "(505.4, 568.45]        4\n",
       "(694.55, 757.6]        2\n",
       "(1009.8, 1072.85]      2\n",
       "(820.65, 883.7]        1\n",
       "(946.75, 1009.8]       1\n",
       "(1072.85, 1135.9]      1\n",
       "(1198.95, 1262.0]      1\n",
       "Name: , dtype: int64"
      ]
     },
     "execution_count": 26,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_tokens = df_eli_short.token_count.value_counts(bins=20).head(17)\n",
    "df_tokens.name = ''\n",
    "df_tokens"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "df_tokens.plot.pie(shadow=True, startangle=90, figsize=(5, 5))\n",
    "plt.title('Figure 1: Token counts', fontsize=16)\n",
    "#plt.show()\n",
    "\n",
    "fig.savefig('./image_files/part1_plot1.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- How about choosing top five categories? "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 28,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(-0.262, 64.05]    315\n",
       "(64.05, 127.1]      56\n",
       "(127.1, 190.15]     47\n",
       "(190.15, 253.2]     19\n",
       "(253.2, 316.25]     13\n",
       "Name: , dtype: int64"
      ]
     },
     "execution_count": 28,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_tokens_5 = df_eli_short.token_count.value_counts(bins=20).head(5)\n",
    "df_tokens_5.name = ''\n",
    "df_tokens_5"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 360x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "df_tokens_5.plot.pie(shadow=True, startangle=50, figsize=(5, 5))\n",
    "plt.title('Figure 2: Token counts (top five categories)', fontsize=16)\n",
    "#plt.show()\n",
    "\n",
    "fig.savefig('./image_files/part1_plot2.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Summary\n",
    "\n",
    "- Now, we know that the length of texts are not distributed well. Most of them are under 65 tokens."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 2. Value count by native language\n",
    "- Maybe we can do some simple analysis, such as native language, to help us look through the data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 30,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Arabic        113\n",
       "Korean         97\n",
       "Chinese        81\n",
       "Japanese       28\n",
       "Thai           15\n",
       "Spanish        15\n",
       "Turkish        12\n",
       "Portuguese      7\n",
       "Other           5\n",
       "Taiwanese       5\n",
       "French          5\n",
       "English         2\n",
       "Italian         2\n",
       "Russian         2\n",
       "Swedish         1\n",
       "Farsi           1\n",
       "Polish          1\n",
       "Name: native_language, dtype: int64"
      ]
     },
     "execution_count": 30,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_native = df_eli_short.native_language.value_counts()\n",
    "df_native"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 31,
   "metadata": {
    "scrolled": true
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAAtUAAAF4CAYAAACb2QXGAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADl0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uIDIuMi4yLCBodHRwOi8vbWF0cGxvdGxpYi5vcmcvhp/UCwAAIABJREFUeJzs3XlYVPXix/HPsKoo4p5brrhvuUYmptbVMlMTL+5ZWu6GmrghmGtpmqnXrVwx18rKtLpqt2uaW5pmZlpu5b4gCqgwLL8/+DE5Ig0yOOdwfb+ep+dhDsM5H8Zm+Mx3vud7LCkpKSkCAAAAkGVuRgcAAAAAcjpKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAwHAPaiEqFrgC4CqUagBO6969uypXrnzP/xo3bixJ+uSTT1S5cmVFRUUZnDZjly5d0rBhw9SoUSM1aNBAoaGhunr1qtGxMs2Mj/GNGzf06quvqlatWmrcuLESEhLS3eeHH37Q4MGDs/24w4YN0+HDhzO8z+zZs/XYY49l63ElqXLlylq0aFGm7z9w4ECNHz8+23MAcC0PowMA+N9Qt25djRgxIt12T09PSdJTTz2lNWvWyNfX19XRMsVqteq1117T7du3NX78eLm5uWn69Onq1auXPvnkE7m5MQaRFZ999pm2bdumt956S2XKlJGXl1e6+3z00Uc6efJkth73yJEj+uKLL9SzZ89s3W92SklJ0bRp07R582Z17drV6DgAnESpBpAtfH19VadOnQy/X7BgQRUsWNCFie7P999/ryNHjuiTTz5R9erVJUn58+dX9+7dtX//ftWvX9/ghDnT9evXlTt3brVv397oKKby559/auLEidq5c6dy5cpldBwA2YChFwAucffUhOTkZM2ZM0eBgYGqXbu2Bg0apKVLl6py5cq2n2nevHm6j8UnTZqk5s2b225XrlxZ8+fPV+vWrdWoUSN99dVXkqSff/5ZL730kmrXrq3HH39cEyZM0K1btzLMV79+fa1atcpWqKW/RtmtVqtdpu7du//t79moUSPt3LlTbdu2VY0aNfTcc89p69attvvca9rBkSNHVLlyZe3evVuSNHLkSA0ePFiLFi1SYGCg6tSpo8GDBys2NlZz5szRE088oUaNGmnixIlKTk6229f27dvVqlUr1apVS127dtXPP/9s9/3Tp0+rf//+euyxx1S/fn0NHz7cbsrIyJEj1b9/fw0bNkx169bVkCFDMvx9N2/erA4dOqhOnTpq2rSpZs6caXu8unfvrtmzZ+vWrVuqXLmyZs+ene7nR44cqfXr1+u3336z+/2vXr2q0NBQNWzYUI899pj69u2rP//8U5KUmJiotm3bqnnz5rp9+7ak1H+jNm3aqF27dtqxY4d69OghSQoKCtLIkSMzzC9Jn376qZo1a6batWurT58+On36tCTp6NGjqly5su3/qTQbNmxQjRo1dO3atb/d79+ZMmWKLl++rFWrVqlQoUJZ3g8A86BUA8gWKSkpSkxMTPdfRmbOnKn58+erS5cumjVrliRp+vTpWTr2nDlz1LVrV40bN07169fX77//rm7duslisWjmzJl64403tGnTJoWEhGS4Dx8fH9WtW1eSlJCQoEOHDmnixImqUKGC3Sj1nDlzFBER8bd54uLiNHr0aHXt2lULFixQgQIFNGTIEEVHR9/X77V9+3Zt3rxZEyZM0PDhw20F9uDBg3rrrbfUvn17RUZGatOmTXY/N378eHXr1s1WcHv27GkrgFeuXFGXLl107tw5TZ06VW+++aYOHDigXr162c13/u9//6v4+Hj961//UnBw8D3zrVmzRgMHDlTNmjU1Z84cdevWTYsXL9aoUaMkSREREQoKClKuXLm0Zs0adezYMd0++vfvr6ZNm6p06dJas2aNqlevrtu3b6tHjx7at2+fwsLCNHXqVF25ckXdunXT9evX5eHhocmTJ+vChQtasGCBJGn+/Pk6efKk3n77bdWuXVvh4eGSUstr//79M3yMb926pXfeeUeDBw/W1KlTderUKb3yyiuyWq2qXLmyqlatqo0bN9r9zIYNG9S0aVMVKFDA0T9hhoYMGaKPP/7Y7k0cgJyN6R8AssV///vfexaEnTt3ppv2ERsbqyVLlqhPnz7q27evJCkwMFBt27bV0aNH7/vYjRs3VpcuXWy3J0+erEKFCmnhwoW2Obxly5ZV165dtXfvXjVo0OBv99erVy/t2bNH3t7emjt3rm3EWpKqVavmMI/VatXw4cP13HPPSZIKFSqktm3bavfu3WrZsmWmf6+bN29q1qxZKlq0qCTp888/1++//66PP/5YefPmVWBgoL788ksdPHhQzz//vO3nRowYYSuwderUUfPmzbVy5UoNGDBAy5YtU3x8vBYvXmz7d6lVq5ZatmypTZs2qV27dpJSR4PHjx+f4ZSd5ORkzZw5U61bt9a4ceMkSU8++aTy5cuniIgI9e7dW1WqVNEjjzwiNze3DKcGPfrooypYsKDOnTtnu8/q1at18uRJbdiwQRUqVJAkBQQEqFmzZoqMjNTAgQNVvXp1vfLKK/rggw9Uu3ZtLViwQIMGDbJ90lGxYkVJkr+/vx599NEMH+O0ec0BAQGSpPLly6tNmzbauHGj2rVrp3bt2mn69OmKiYlRvnz5FBUVpR07dujdd9918K/39/z9/Z36eQDmw0g1gGxRr149ffTRR+n+u9eJiQcPHlRCQoKefvpp2zaLxaJ//OMfWTp2WvFKs3v3bjVu3Fhubm62EfM6deoob9682rlzp8P9vf7661qyZImeeeYZ9enTR9999919Z7qzRD7yyCOS9LfTT+6lePHitkItpZbzcuXKKW/evLZtfn5+iomJsfu5O4t7wYIFVadOHR06dEhS6mNTp04d+fr62h6b4sWLq0KFCnaPjaM58MePH1dUVJRatWpltz2t3P/www/39bveaffu3SpTpozKlCljy5grVy7Vq1dPu3btst1v0KBBKlGihPr166fq1aurd+/e932sfPny2Qq1lFp2S5cubXu82rRpo+TkZG3evFmStGnTJvn4+Oipp57K8u8H4H8TI9UAskW+fPlUs2bNTN03bSrC3aWtcOHCWTr23XNSo6OjtWbNGq1ZsybdfS9fvuxwf2nTPQICAnTq1Cl98MEHatKkyX1luvPks7SVQ+6e++yIj49Pum25c+f+25/x9PRM90amYMGCOnXqlKTUx+bgwYP3/FShSJEitq8dzfO9fv36Pe+XN29eeXt7KzY29m9//u9ER0frxIkT98xYtmxZ29fe3t5q2bKlFixYoMaNG8vd3f2+j3Wv37NgwYK6dOmS7ftNmjTRxo0b9eKLL2rDhg1q1arVPVcxAfBwo1QDcLm00deoqCgVK1bMtv1e6yvfXURv3rzpcP958+ZVixYt1Llz53Tfy2ge7NGjR/Xbb7/ZTaOwWCyqUqWK9u3b5/CY98NisaT7veLi4rJl31arVbdu3bIr31euXJGfn58k2aaN3Gtd6HuV+Iyk7e/udbxv3Lih+Ph42/ezIl++fKpSpYomTpyY7nt3ltnTp09r2bJlqly5sj744AO1adNG5cqVu69j3bhxI922K1euqFKlSrbbbdu21RtvvKFjx47pwIEDCg0Nva9jAHg4MP0DgMtVrVpVPj4+ditiSNI333xjdztv3ry2EUMptWD/+OOPDvdfr149nThxQjVq1FDNmjVVs2ZNFS9eXNOnT9dvv/12z5/Zv3+/3njjDf3xxx+2bQkJCfrhhx/sClZ2yJs3r27fvm1X6LKzuN85XeXSpUs6cOCAGjZsKOmvx6Zy5cq2x6ZSpUqaM2fOfWUoV66cChQokG5ljLSTJtNO+syMu9cAr1u3rs6cOaOSJUvaMtaoUUNLly7Vt99+Kyl1LnRYWJhKliyp1atXq2TJkgoLC7NdQTGzo9ZRUVF2F4g5fPiwzpw5Y3u8JKlFixbKkyeP3nzzTZUqVUr16tXL9O8G4OHBSDUAl8uXL59eeuklLViwQF5eXqpatao+++wzHT58WBaLxXa/wMBALVmyRJGRkapYsaJWr16tq1evOhxR7d+/vzp16qTXX39dHTp0UEJCgubOnavz589neKLh888/r0WLFmnAgAEaNGiQPDw8tHTpUl26dMm2Ookk/fLLL/Ly8rKdCJcVTZo00ZQpUzRmzBh17dpVv/76q1auXJnl/d3Jzc1NkydP1q1bt+Tj46M5c+bIz89PnTp1kiS9/PLL+uyzz9S7d2/16NFDnp6eWrx4sQ4cOPC3q6Pczd3dXQMHDtSECROUP39+tWjRQkePHtXs2bPVqlWr+3oj4uvrqwsXLmjHjh2qUaOGgoKCFBkZqVdeeUWvvfaa/Pz8tGbNGv373//WCy+8ICl15ZE9e/YoMjJSefLkUXh4uHr27KmVK1eqa9euypcvn6TUE2jz5MmTbt59Gi8vLw0dOlRvvPGGrFar3nnnHVWpUsVuXrqXl5eeffZZrVmzRgMGDMjU7/TDDz/cs9gHBwc7nMIDIGeiVAMwxMCBA5WUlKRly5YpNjZWTz31lLp06aL169fb7tO3b19dvnxZ7777rjw8PPTCCy+oT58+WrFixd/uu0aNGlq2bJlmzpypwYMHy9vbW3Xr1tXUqVPtppvcKV++fFq+fLmmTp2q8PBwxcfHq27dulq5cqXd2tkDBw5UyZIlFRkZmeXfvUKFCpo4caLmzZunV199VbVr19asWbP0z3/+M8v7TOPh4aGwsDDbOsgNGjTQrFmzbNMxSpQooZUrV2ratGkaPny4LBaLqlevriVLlqhq1ar3daxu3bopV65cWrx4sdatW6eiRYvq5Zdf/tsl7O4lODhY//nPf9SnTx9NnTpVzz33nD788ENNnTpV48aNU0JCgvz9/TV37lw1bdpUFy9e1LRp09S2bVvbiHJAQICef/55TZ8+Xc2aNZO/v7/atm2rBQsW6Oeff9b8+fPveeySJUuqZ8+eevPNNxUXF6emTZtq7Nix6eZMBwYGas2aNbZS78g333yT7pMXSWrdujWlGvgfZUlJ+6wMAFwkISFBmzZt0pNPPml3cuKwYcN04sQJu2INmMG4ceN09OhRrVq1yugoAEyKkWoALufl5aW5c+dq3bp16t27t3Lnzq2dO3dq06ZN9zw5DTDKRx99pCNHjmjt2rWaMWOG0XEAmBgj1QAMcfLkSb3zzjvat2+fbt68qXLlyqlnz55q37690dEAm3Hjxumzzz5Tx44dNXr0aKPjADAxSjUAAADgJJbUAwAAAJyUI+dUJycnKykp+wbY3d0t2bq/7ES2rDNzPrJlnZnzmTmbZO58ZMs6M+czczbJ3PnIlnXZmc/TM/NXas2RpTopKUXR0Y6vqpZZfn55snV/2YlsWWfmfGTLOjPnM3M2ydz5yJZ1Zs5n5mySufORLeuyM1+RIvkyfV+mfwAAAABOolQDAAAATqJUAwAAAE6iVAMAAABOolQDAAAATqJUAwAAAE6iVAMAAABOolQDAAAATqJUAwAAAE6iVAMAAABOolQDAAAATvIwOsCDlNc3t3J7Z+5XdHRt91vxiYq9cSs7YgEAAOB/zP90qc7t7aGyIzdmy75OvdVasdmyJwAAAPyvYfoHAAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOCkB1qqDx48qO7du0uSTp8+rc6dO6tLly6KiIhQcnKyJGnOnDkKCgpSp06d9NNPPz3IOAAAAMAD8cBK9fvvv6+wsDDFx8dLkqZMmaKQkBCtXLlSKSkp2rp1qw4fPqw9e/Zo3bp1mjFjht58880HFQcAAAB4YDwe1I4fffRRzZ49W6GhoZKkw4cPq2HDhpKkwMBA7dixQ+XKldOTTz4pi8WiEiVKKCkpSVFRUSpYsODf7tvd3SI/vzwPKnqGjDimu7ubIcfNDDNnk8ydj2xZZ+Z8Zs4mmTsf2bLOzPnMnE0ydz6yZZ1R+R5YqW7ZsqXOnDlju52SkiKLxSJJ8vHxUUxMjGJjY+Xn52e7T9p2R6U6KSlF0dE3HWYoUiRfFtPfW2aOmd38/PIYctzMMHM2ydz5yJZ1Zs5n5mySufORLevMnM/M2SRz5yNb1mVnvvvpki47UdHN7a9DxcXFydfXV3nz5lVcXJzd9nz5srcIAwAAAA+ay0p1tWrVtHv3bknStm3bVL9+fdWtW1fbt29XcnKyzp07p+TkZIej1AAAAIDZPLDpH3cbMWKExo4dqxkzZqh8+fJq2bKl3N3dVb9+fQUHBys5OVnh4eGuigMAAABkmwdaqkuVKqW1a9dKksqVK6cVK1aku8+gQYM0aNCgBxkDAAAAeKC4+AsAAADgJJdN/4C9Qvk95OaVO1P3zcyZp8kJt3T1eqKzsQAAAJAFlGqDuHnllsblz779jbsuKSbb9gcAAIDMY/oHAAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CQPVx7MarVq5MiROnv2rNzc3DRhwgR5eHho5MiRslgs8vf3V0REhNzc6PoAAADIOVxaqv/73/8qMTFRq1ev1o4dOzRz5kxZrVaFhISoUaNGCg8P19atW/XMM8+4MhbuktfPU7k9c2XqvkWK5HN4n1vW24qNtjobCwAAwLRcWqrLlSunpKQkJScnKzY2Vh4eHjpw4IAaNmwoSQoMDNSOHTso1QbL7ZlLNZfVzLb9HXrpkGJFqQYAAP+7XFqq8+TJo7Nnz+rZZ5/VtWvXNH/+fO3du1cWi0WS5OPjo5iYGIf7cXe3yM8vz4OOm44Rx7wfZs5nRDZ3dzfTPiZkyzoz5zNzNsnc+ciWdWbOZ+ZskrnzkS3rjMrn0lK9dOlSPfnkkxo2bJjOnz+vl156SVbrXyOYcXFx8vX1dbifpKQURUffdHi/zExNuB+ZOWZmZXc2KfvymTnb/fDzy2PIcTODbFln5nxmziaZOx/Zss7M+cycTTJ3PrJlXXbmu59O5NIzAn19fZUvX2q4/PnzKzExUdWqVdPu3bslSdu2bVP9+vVdGQkAAABwmktHqnv27KnRo0erS5cuslqtGjJkiGrUqKGxY8dqxowZKl++vFq2bOnKSAAAAIDTXFqqfXx89N5776XbvmLFClfGAAAAALIVC0IDAAAATqJUAwAAAE6iVAMAAABOolQDAAAATqJUAwAAAE5yWKpHjx6dbtvgwYMfSBgAAAAgJ8pwSb2IiAhdvHhR+/btU1RUlG17YmKi/vzzT5eEAwAAAHKCDEt1UFCQfvvtNx09etTugizu7u6qU6eOS8IBAAAAOUGGpbpmzZqqWbOmAgICVLx4cbvv/frrrw88GAAAAJBTOJxT3blzZ+3bt892e/ny5erZs+eDzAQAAADkKA4vUz558mQNHTpUnTp10sGDBxUTE6N169a5IhsAAACQIzgs1U888YTCw8M1cOBAFS5cWB9//LGKFi3qimwAAABAjuBw+se0adMUERGhuXPnqnv37urQoYM2b97simwAAABAjuBwpPrw4cNav369ihQpombNmunxxx/XsGHD9Mwzz7giHwAAAGB6Dkv1kiVLZLFYdOPGDfn6+qpWrVr69NNPXZENAAAAyBEcTv84deqUnnvuObVu3VoXL17Us88+qwsXLrgiGwAAAJAjOCzVEyZM0JgxY1SoUCEVK1ZM3bp1U3h4uCuyAQAAADmCw1IdHR2txo0b22537dpVsbGxDzQUAAAAkJM4LNWSFB8fL4vFIkm6fPmykpOTH2goAAAAICdxeKJi586d1atXL129elXTp0/Xxo0b1bt3b1dkAwAAAHIEh6W6Y8eOKlu2rL799lslJiZqwoQJdtNBAAAAgIedw1I9c+ZMhYSEqEGDBrZtEydOVFhY2AMNBgAAAOQUGZbqWbNm6caNG9q0aZPdiYlWq1Xbt2+nVAMAAAD/L8NSXbt2bR06dEhubm7y8/OzbXd3d9c777zjknAAAABATpBhqW7atKmaNm2qwMBA1apVy5WZAAAAgBzF4ZJ6FGoAAADg72VqnWoAAAAAGcuwVG/ZskWSlJCQ4LIwAAAAQE6UYal+7733JEnBwcEuCwMAAADkRBmeqOjj46OWLVvq4sWLatOmTbrvb9iw4YEGAwAAAHKKDEv1Bx98oCNHjmjMmDEaO3asKzMBAAAAOUqGpTpv3rxq0KCBFixYoKJFi+rw4cNKTExUrVq1lDdvXldmBAAAAEzN4WXKY2Ji1L17dxUuXFhJSUm6ePGi5s+fr7p167oiHwAAAGB6Dkv122+/rXfeeUePP/64JGnnzp166623tHbt2gceDgAAAMgJHK5THRcXZyvUkhQQEKBbt2490FAAAABATuKwVFssFp09e9Z2+8yZM3J3d3+goQAAAICcxOH0jwEDBig4OFgBAQGyWCzavn27IiIiXJENAAAAyBEcluqnn35a5cuX165du5ScnKw+ffqoQoUKrsgGAAAA5AgOS7UklS9fXuXLl8+WAy5YsEDffPONrFarOnfurIYNG2rkyJGyWCzy9/dXRESE3NwczkoBAAAATMOl7XX37t368ccftWrVKkVGRurChQuaMmWKQkJCtHLlSqWkpGjr1q2ujAQAAAA4zaWlevv27apUqZIGDBigvn376qmnntLhw4fVsGFDSVJgYKC+//57V0YCAAAAnOZw+kdoaKimTp2aLQe7du2azp07p/nz5+vMmTPq16+fUlJSZLFYJEk+Pj6KiYlxuB93d4v8/PJkS6b7YcQx74eZ8xmRzd3dzbSPCdmyzsz5zJxNMnc+smWdmfOZOZtk7nxkyzqj8jks1UeOHLErvs7w8/NT+fLl5eXlpfLly8vb21sXLlywfT8uLk6+vr4O95OUlKLo6JsO71ekSD6n8t4tM8fMrOzOJmVfPjNnux9+fnkMOW5mkC3rzJzPzNkkc+cjW9aZOZ+Zs0nmzke2rMvOfPfTiRyW6qJFi6p169aqXbu2fHx8bNvDwsLuO1i9evW0fPlyvfzyy7p06ZJu3bqlgIAA7d69W40aNdK2bdvsLjQDAAAA5AQOS/Vjjz2mxx57LFsO1qxZM+3du1dBQUFKSUlReHi4SpUqpbFjx2rGjBkqX768WrZsmS3HAgAAAFzFYakeOHCgbt++rdOnT8vf31/x8fHKnTt3lg8YGhqabtuKFSuyvD8AAADAaA5X/zh48KCefvpp9enTR5cuXdJTTz2l/fv3uyIbAAAAkCM4LNVvv/22li5dKj8/Pz3yyCOaOnWqJk2a5IpsAAAAQI7gsFTfvn1bFStWtN1u2rSpkpKSHmgoAAAAICdxWKo9PDx0/fp125J6J06ceOChAAAAgJzE4YmK/fr1U7du3XT58mUNHTpUO3bs0Pjx412RDQAAAMgRHJbqZs2aqXz58tqxY4eSk5M1YMAAVahQwRXZAAAAgBzB4fQPSUpMTFRycrI8PDzk4eGwhwMAAAAPFYel+uOPP1aPHj106NAh/fDDD+ratau+/vprV2QDAAAAcgSHw85Lly7V+vXrVbRoUUnSuXPn1KdPH658CAAAAPw/hyPVnp6etkItSSVKlJCnp+cDDQUAAADkJBmOVB8+fFiSVLlyZY0fP17BwcFyd3fXJ598orp167osIAAAAGB2GZbqQYMG2d3+9ttvbV9bLBaFhYU9sFAAAABATpJhqf7mm29cmQMAAADIsRyeqHj58mWtX79e0dHRdttDQ0MfWCgAAAAgJ3F4omK/fv30008/KSUlxe4/AAAAAKkcjlRbrVbNmTPHFVkAAACAHMnhSHX16tV17NgxV2QBAAAAciSHI9V169ZVu3btVKRIEbtLlG/duvWBBgMAAAByCoeletGiRXrnnXf06KOPuiIPAAAAkOM4LNW+vr567rnnXJEFAAAAyJEclurHH39cb7/9tv7xj3/Iy8vLtr169eoPNBgAAACQUzgs1Rs2bJAkff3117ZtFouFOdUAAADA/3NYqrmyIgAAAPD3HJbqJUuW3HP7yy+/nO1hAAAAgJzIYam+c43qhIQE7d27VwEBAQ80FAAAAJCTOCzVU6ZMsbt98eJFjRkz5oEFAgAAAHIah1dUvFuxYsV09uzZB5EFAAAAyJHua051SkqKfv75ZxUqVOiBhgIAAABykvuaUy1JxYsXV2ho6AMLBAAAAOQ09z2nGgAAAIC9DEv1qFGjMvwhi8WiyZMnP5BAAAAAQE6TYan29/dPt+3atWtatmyZSpYs+UBDAQAAADlJhqX6lVdesbv9/fffa8SIEWrTpo3CwsIeeDAAAAAgp3A4pzoxMVHTp0/X+vXr9eabb6ply5auyAUAAADkGH9bqk+dOqWhQ4fKx8dHn376qR555BFX5QIAAAByjAwv/vLxxx/rn//8p5555hlFRkZSqAEAAIAMZDhSPWbMGLm5uWnhwoV6//33bdtTUlJksVi0f/9+lwQEAAAAzC7DUr1161ZX5gAAAAByrAxL9YNcNu/q1at68cUXtXjxYnl4eGjkyJGyWCzy9/dXRESE3NwynJUCAAAAmI7L26vValV4eLhy5colKfWKjSEhIVq5cqVSUlIYIQcAAECO4/JS/fbbb6tTp04qWrSoJOnw4cNq2LChJCkwMFDff/+9qyMBAAAATnG4TnV2+uSzDX7lAAAgAElEQVSTT1SwYEE1adJECxculPTXiY+S5OPjo5iYGIf7cXe3yM8vzwPNei9GHPN+mDmfEdnc3d1M+5iQLevMnM/M2SRz5yNb1pk5n5mzSebOR7asMyqfS0v1xx9/LIvFop07d+rIkSMaMWKEoqKibN+Pi4uTr6+vw/0kJaUoOvqmw/sVKZLPqbx3y8wxMyu7s0nZl8/M2e6Hn18eQ46bGWTLOjPnM3M2ydz5yJZ1Zs5n5mySufORLeuyM9/9dCKXluoPP/zQ9nX37t01btw4TZs2Tbt371ajRo20bds2Pf74466MBAAAADjN8GU2RowYodmzZys4OFhWq5XLoAMAACDHcelI9Z0iIyNtX69YscKoGAAAAIDTDB+pBgAAAHI6SjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkSjUAAADgJEo1AAAA4CRKNQAAAOAkD6MDAPerQF5PeeTOlan7FimSz+F9Em/d1rVYq7OxAADAQ4xSjRzHI3cuHalSNdv2V/XXIxKlGgAAOIHpHwAAAICTKNUAAACAkyjVAAAAgJMo1QAAAICTKNUAAACAkyjVAAAAgJMo1QAAAICTKNUAAACAkyjVAAAAgJMo1QAAAICTKNUAAACAkyjVAAAAgJMo1QAAAICTKNUAAACAkzxceTCr1arRo0fr7NmzSkhIUL9+/VSxYkWNHDlSFotF/v7+ioiIkJsbXR8AAAA5h0tL9eeffy4/Pz9NmzZN165dU/v27VWlShWFhISoUaNGCg8P19atW/XMM8+4MhYAAADgFJeW6latWqlly5a22+7u7jp8+LAaNmwoSQoMDNSOHTsclmp3d4v8/PI80Kz3YsQx74eZ85k5m+T6fO7ubqZ9TMycTTJ3PjNnk8ydj2xZZ+Z8Zs4mmTsf2bLOqHwuLdU+Pj6SpNjYWA0ePFghISF6++23ZbFYbN+PiYlxuJ+kpBRFR990eL8iRfI5F/gumTlmZmV3Nin78pk5m2T+fJnh55fH5cfMLDNnk8ydz8zZJHPnI1vWmTmfmbNJ5s5HtqzLznz30zlcPnn5/Pnz6tGjh9q2bas2bdrYzZ+Oi4uTr6+vqyMBAAAATnFpqb5y5YpeeeUVDR8+XEFBQZKkatWqaffu3ZKkbdu2qX79+q6MBAAAADjNpaV6/vz5unHjhubOnavu3bure/fuCgkJ0ezZsxUcHCyr1Wo35xoAAADICVw6pzosLExhYWHptq9YscKVMQAAAIBsxYLQAAAAgJMo1QAAAICTKNUAAACAkyjVAAAAgJMo1QAAAICTKNUAAACAk1y6pB7wvy6/b255eWfuaZWZS58mxCfq+o1bzsayyZ/PS165vB3eL1PZbsfrekxCdsQCACDHo1QD2cjL20P/6vtNtu1vwPzm2bYvSfLK5a3pwc9ny76GrflColQDACCJ6R8AAACA0yjVAAAAgJMo1QAAAICTKNUAAACAkyjVAAAAgJNY/QOAKRT0zS33bFyOMCk+UVHZuBwhAAB/h1INwBTcvT10ZuR32ba/Um81ybZ9AQDgCNM/AAAAACdRqgEAAAAnUaoBAAAAJ1GqAQAAACdRqgEAAAAnUaoBAAAAJ1GqAQAAACdRqgEAAAAnUaoBAAAAJ3FFRQBwIH9+b3l5eWXqvpm5hHpCQoKuX493NhYAwEQo1QDggJeXl8aNG5dt+0vdF6UaAP6XMP0DAAAAcBKlGgAAAHASpRoAAABwEnOqASCH8/PzkKdn7kzd19GJlFbrLUVHJ2ZHLEmSj18e5fF0z9R9M3OS501rkuKibzobyyavb27l9nb8pzAz2W7FJyr2xq3siAUgB6JUA0AO5+mZW1u/qZAt+2rR/LikmGzZlyTl8XTXI/85kG37u9CsjuKybW9Sbm8PlR25MVv2deqt1orNlj0ByImY/gEAAAA4iVINAAAAOIlSDQAAADiJUg0AAAA4iVINAAAAOInVPwAAMKHMLvcnuX7Jv0L5PeTmlT3LOEpScsItXb2efUs55vXzVG7PXJm6r6N8t6y3FRttzY5Y+B9HqQYAwISyc7k/KXuX/HPzyi2Ny59Ne5Pcxl1Xdi7lmNszl2ouq5kt+zr00iHFilINx0xRqpOTkzVu3DgdPXpUXl5emjhxosqUKWN0LAAAgGxVIK+nPHJnzyi6JCXeuq1rsdlT+vP75pZXNn46khCfqOvZeEGk/Pm85JXLO1P3zVS+2/G6HpPgbCwbU5TqLVu2KCEhQWvWrNGBAwf01ltvad68eUbHAgAAyFYeuXPpSJWq2ba/qr8ekbKpVHt5e+hffb/Jln1J0oD5zbNtX5Lklctb04Ofz7b9DVvzhZSNpdqSkpKSkm17y6IpU6aoVq1aat26tSSpSZMm+u677wxOBQAAAGSOKVb/iI2NVd68eW233d3dlZiYfScsAAAAAA+SKUp13rx5FRcXZ7udnJwsDw9TzEwBAAAAHDJFqa5bt662bdsmSTpw4IAqVapkcCIAAAAg80wxpzpt9Y9jx44pJSVFkydPVoUKFYyOBQAAAGSKKUo1AAAAkJOZYvoHAAAAkJNRqgEAAAAnPbSl+tat1Cv8XLp0yeAkAAAAyOkeylI9Z84cvffee5KkiRMnauHChQYnAoAHg9NmAMA1HsoTFV988UV98sknttudOnXS6tWrDUxk78iRI1qzZo3i4+Nt26ZMmWJgInuxsbE6e/asSpcurTx58hgdx+bYsWMaN26cYmJi1KZNG/n7+6tZs2ZGx8oRDh06pPXr19s+wZHM9f9cTEyMduzYodu3b9u2tWvXzsBE6V2/fl358+c3OkY6r7zyihYvXmx0jAzdvHlT586dU6lSpZQrVy6j49g5c+aMvv76a7vnxcCBAw1M9Jec8Jy4evWq3d+xEiVKGJgmZ0hKStIvv/xi9+/aoEEDAxPlHN9++61Wrlxp99gtX77cpRkeyiusWCwWJSQkyMvLS1ar1XQjOSNHjlS3bt30yCOPGB0lna+++krz589XUlKSWrVqJYvFov79+xsdS5I0adIkTZkyRWFhYQoKClLv3r1NUarHjx+v8PBwBQcHy2KxSEodPbRYLKZ5Mzdu3Dh169ZNhQsXNjrKPQ0YMEAlS5a05Ut7HM1gz549Gj9+vO05UaJECXXs2NHoWDb58uXTli1bVK5cObm5pX44Wa5cOYNTpdq8ebNmz56t5ORktWrVSp6enurTp4/RsWyGDRumJk2amPJ5YebnhJT6mrJt2zYVLVrUdK93O3bs0JIlS5SQkGDb5urylZHBgwfrxo0bKlKkiKTUf1czlOqc8Hfsvffe06hRowx9vj6UpbpTp05q06aNKlWqpBMnTqh3795GR7JTuHBhU/1RvtPSpUu1du1a9erVS/3791eHDh1MU6olqUyZMrJYLCpYsKB8fHyMjiNJtsdnxowZBifJWN68edW+fXujY2QoJSXFVCPnd3rvvfe0YsUKDRo0SH379lXnzp1N9fyNiorSsmXLbLctFotpCsSiRYu0bt069e7dW/3791dQUJCpSnWuXLlMMzJ9NzM/JyTpp59+0pYtW2xv5MxkypQpGj16tCkHrq5du6aVK1caHSOdnPB3LH/+/GrYsKGhGR7KUt2xY0e1aNFCf/75p0qXLq2CBQsaHclOyZIltXDhQlWtWtX2jvDJJ580OFUqNzc3eXl5yWKxyGKxKHfu3EZHssmfP79Wr16tW7duaePGjfL19TU6kiTZ3jUnJibqq6++ktVqlZR6kuz48eONjKbt27dLSh3NnD9/vqpXr26q/+fSRpJKly6tH3/8UdWrV7d9z8vLy6hYdtzc3OTn5yeLxSJvb2/TvJlLExkZqZiYGNuULTPlc3d3l7e3tywWi9zc3EzzenLy5ElJqc/dDRs22D0vjB7lzwnPCSl1gCM+Pt40/6Z3Kl68uJ544gmjY9xTiRIldP78eRUvXtzoKHbS/o7dvHlTsbGxcnNz04wZM9S3b1+VLFnS0Gxr1qyRJHl6emrs2LF2z9fg4GCXZnmoSvXcuXPVv39/DR06NN1HZdOnTzcoVXpWq1UnT560vbBL5ig4klS/fn0NGzZMFy9eVHh4uGrWrGl0JJvJkydr/vz5KlCggH7++WdNmjTJ6Eh2RowYoWbNmmn//v0qWrSobt68aXQkbdy4UVJqqT59+rROnz5t+54Z/p9Lm2KUkpKiXbt22bZbLBZt3brVwGR/efTRRzV9+nRFR0dr4cKFpps3+vXXX2vevHmmnLJVp04dDR8+XBcvXtT48eNVrVo1oyNJksLDw21fr1271va1GUb5c8JzQpLOnz+vZs2aqUyZMpJkqmkChQoVUnh4uKpVq2ZY+bpb2uttQkKCvvrqK/n5+dm+lzb4YQYREREaM2aMZs+erSFDhmjatGkKCAgwNNPly5clSbVr15YkXblyxbAsD9WJir/++quqVKmiPXv2SJLthUmS4R8Z/J1Lly6paNGiRsew2bZtm44dO6YKFSqYYs7ynWJiYmSxWLRlyxY1a9bMVCeO9ezZU0uXLtWoUaM0ZcoUdenSxVQf8yUlJSklJUUHDhxQrVq1TDXqdaekpCS5u7sbHcMmMTFR69atsz0n/vnPf5rqsevUqZOWL1+uXr16afny5erQoYPdidpG+89//mN77J5++mmj49iJj4/X8ePHVa1aNW3ZskVNmzaVp6en0bHSMdtzQpLOnj2bbpvRI5pp5syZk26bWaf5mE2PHj30wQcfqF+/flq0aJHpFnqIjY2VJMM6wEM1Ul2lShVJkr+/vxYsWKBTp07J399fr732msHJ7M2aNUsrV66U1WrV7du3VbZsWduIotEuXbqkEiVKqFSpUvrggw/0yCOPqGrVqkbHkiSFhoaqcePG+vHHH5WcnKzNmzfrX//6l9GxbFJSUnT58mXdvHlTN2/e1PXr142OZDNt2jSVLl1a586d0+HDh1WkSBG99dZbRsey+fLLL5WcnKyEhARNmzZNvXr1Uq9evYyOJSl1ZKlZs2Z6+umntXbtWl2+fNk05UEy95StK1eu6NFHH1W5cuW0ePFilSpVyvY6bQbDhw9XQECAqlWrppMnT+rLL780zaeaZn5OSOac7pZm4MCBunTpkhITE5WSkmKq61Xs3btXt27dUkpKiiZMmKDXX39dbdq0MTqWjcVi0bBhwxQYGKhNmzaZ6vXEDB3AfGcQuMCIESP06KOPKiQkRMWKFdOIESOMjmRn27Zt2rZtm9q0aaNNmzapWLFiRkeyGTFihK5cuaKZM2eqcePGmjx5stGRbM6ePau2bdvq+PHjGj9+vO0dq1kMHDhQmzdv1gsvvKAWLVooMDDQ6Eg2+/btU6dOnfTjjz9q0aJFOn/+vNGR7CxevFhPPPGEPv/8c3377bf6z3/+Y3QkmzfeeEOHDx/WtGnT5OnpaTd1wAzq16+voUOHmnLKVmhoqC5duqQZM2aoXr16pno9kaSLFy+qc+fOkqRXX33VVOXLzM8JSba/q/v379eZM2cUHR1tcKK/jB49Wj179lTXrl0VFBRkqhM+p02bprJly2r58uVatWqVqUaBJendd99VUFCQevTooYIFC+rdd981OpKNGTrAQ1mq4+Pj1aVLF1WpUkXdunVTTEyM0ZHs+Pn5ycvLS3FxcSpTpozdGqlGS0xMVIMGDXTjxg21bt1aycnJRkeysVqt2rRpkypWrKioqChTvYhLqWuNdunSRS1atNDOnTtN9WYuOTlZP/30k0qVKqWEhARFRUUZHcmOt7e3JMnHx8f23DCLGzduqHnz5rpw4YJee+01u2W6zGDo0KFq166dOnbsqGbNmmnkyJFGR7KxWq1q1KiRbty4obZt2yopKcnoSOmkndvyxx9/mOr1zszPCSl15ZQ+ffqoWLFieuuttwyd53q3EydOaOPGjXryySe1ceNG22NpBt7e3ipUqJA8PDxUpEgR072eDBgwQE2bNpXFYtHjjz9uN/fbaGboAA/V9I+0F8cCBQroyy+/VP369W1FwkweeeQRffTRR8qdO7emT59uqhFXq9WqKVOmqH79+tq1a5ep/gj27t1bmzZt0siRIxUZGamQkBCjI9mZM2eOVqxYIQ+Pv552ZjkBpW3btpowYYImT56sadOmqUePHkZHslOqVCl16NBBY8eO1Zw5c1SrVi2jI9lYrVYtXrxY1atX1++//266cnPmzBn99ttvun37tg4fPqzDhw+bZv6o1WrV1KlTVb9+fe3du1eJiYlGR7IzevRohYSE6OrVqypatKjefPNNoyPZmPk5IZl7upuPj48sFotu3rypggUL2qaomIGPj49efvlldenSRR9++KHpVgHJnz+/li1bZrfuvRlOapdSO8DGjRs1atQowzrAQ3WiYvfu3e+53QxndN8pOTlZFy5ckK+vr9avX6+AgABVrFjR6FiSpFOnTmnHjh3q2LGjtmzZopo1a6p06dJGx7I5efKk/vjjD1WuXFnFihUz1QUROnTooA8//NB0V43LKeLi4uTj46MrV66Y6mIc+/fv15YtW9S3b19t2LBBNWvWNFXBCQ4OTncBk06dOhmY6C8nTpzQ9u3b1alTJ23evFnVq1dX2bJljY6VY5j1OSGlzg3+7bffVKxYMYWFhaldu3am+XRuxowZyp8/v65cuaILFy7ozz//1EcffWR0LEmp52j88ccfqlixoo4dO6ayZcua6sTnUaNGpdtm9PSZxMREeXh43HNU39WP3UNVqjNitVpNdUZ3bGys3n//fV2+fFlPPfWUKleubFuWyGhJSUn65JNPdP78eTVq1Ej+/v6mWed7xYoV2rx5s65fv6727dvr9OnTpprf+tprr2nu3Ll2I9VGGzx4sGbNmnXPkQYzjKLnlGUwv//+e505c0a1atVSuXLlTPVx8ksvvWR38RczSU5O1qeffqpz587p8ccfV8WKFU31cfKnn36qhQsX2l1q2+hl63LKc8Ls4uLi5O3trW3btql27doqVKiQoXnWrVunjh07avr06en+XYcOHWpQKsfMsDrZsGHDNH36dDVv3txuVTcjlpk0z193F1q9erWWLFliO/PX09NTX3/9tdGxbEaPHq3AwEDt3btXhQsX1pgxY7RixQqjY0lKXb+1aNGi+v7771WjRg2NGDFC77//vtGxJKWuubxy5Ur16NFDL730kjp06GB0JEmpT3gpdaWD9u3by9/f3/aiafQfwVmzZkkyR4G+l+bNm0syz8jqvcyYMUMXLlzQ8ePH5enpqYULF5riqmN3XsDkiy++sFuT1+gLmKSJiIhQgQIFtGvXLlWpUkWhoaFauHCh0bFs3n//fc2bN89UH8Gb/Tlh5jfqGRXXAwcOGF5c067uWL58eUNzOGLG1cnS/o5+8803huaQHtJSvXbtWkVGRmrevHlq1aqV6UZxoqOjFRQUpM8//1x169aVmT5M+OOPPzRp0iTt27dPzZs3N9UfwDvfnUrmubrY999/r/fee8/oGH/ryJEjWrNmjd2InNEf6Ul/LYN56dIlPf/887avR48ebZq15fft26cPP/xQ3bt3V/v27bVq1SqjI0myv4BJ2hXHJHNNdzt16pQmTJigHj166Omnn9YHH3xgdCQ7pUuXNs2nhGkOHjyogwcP3vN7ZnhOmPmNupmLa65cubR3717TneN1t7TVySZPnqyXX37ZFOcZBAcHZzjV09WrpzyUpbpAgQIqWrSo4uLi1KhRI9uLgJkcP35cknThwgXbyQBmkJSUZFsZIu1SpWbx/PPPq2vXrjp37pxeffVV01xIomLFiqb4Y/d3Ro4cqW7dutn+6JjNp59+Kh8fHyUkJGjGjBkaPHiw0ZFskpKSFB8fL4vFoqSkJNM8JyIjIyWlXlzlzos0bdq0yahI6SQmJtrO0L9586ZpHrs0uXLlUu/evVW1alXbH22jRzTTrh5nVvealpLG6E/mmjRpIklq2bKlbty4IXd3d61du1bt2rUzNJck25vxP/74Q1arVTVr1tQvv/wiHx8f23PZDMy4OpkZPhlM81CW6nz58mnLli22y6aabfmwsLAwjR49WsePH9fgwYMVERFhdCSbkJAQde7cWZcvX1ZwcLBGjx5tdCSbbt26KSAgQMeOHVO5cuVMcxGJM2fOZPikN/oPdJrChQurY8eORsfI0Jw5c9S3b1/Fx8dr1apVppnHL6XOWX7xxRcVFRWljh07qmfPnkZHkpRapvfv36+NGzfqxx9/lJT6BuCbb77Rc889Z3C6VIMHD1ZwcLCuXLmijh07muZEtjRNmzY1OkI6aSu37Nmzx668enh46MKFC4a/MTbrtJQ7vfHGG3rxxRf173//WxUrVlR4eLgWLVpkaKa0vxF3nnuTlJRkuovTmXF1srSLbV24cEGTJ0/W8ePHVbZs2XueVPmgPZQnKsbGxuqPP/5Q4cKFtXjxYjVr1kyNGjUyOpbNokWLTHVlrHuJiopSgQIFTLW6xvnz5/XFF1/YTWEww9Jhzz77bIYvjO3bt3dxmnsLDw9XqVKl7EbkzLBM0p2jXnFxcdqzZ49t1NXoUa87Xb9+XadPn1apUqVMU/jPnz+vXbt2aezYsZo4caKSk5Pl7u6uSpUqmeYqqNJfS68VLlzYdCPViYmJWr9+vSlPzO7atauuXLmi6tWr65dffpGnp6cSEhLUsWNH9e7d2+h4io6O1vbt2+2uWtinTx+jY0lKHYCJjIzUSy+9pOXLl6t79+6mGQ1u37691q1bZ1vNIjg4WOvXrzc6lu0E2eTkZB08eFD+/v6mW52sd+/e6ty5sxo0aKA9e/YoMjLS5dN7H8qR6sGDB2vx4sWSZKoLIaT573//q549e8rd3d3oKOns2LFDS5cutSuuZpmf+frrrysgIMBUJxVJqaPAZinPGbFarTp58qTt5DbJHKU6bdTrxo0b8vX11SuvvGJwovTMuMSUJPn6+urrr79WlSpV9N133+nUqVMqVKiQWrRoYXQ0m507d2rZsmV2S2GlvTabQUREhGlPzM6VK5c+//xzeXt7KyEhQYMGDdLs2bPVrVs3U5TqwYMHq2zZsjp27Ji8vb1NdTlrM68tHxQUpNatW6tSpUr6/fffNWjQIKMjSZJ27dql/v37y83NTe+++67tzYiZxMfH217fnn76aS1ZssTlGR7KUp02/ePOxcvNcja8JF27dk1NmjRRqVKlZLFYbNNUzGDKlCkaPXq04R8x3ouPj4+GDBlidIx0atSoYXQEh+4ugWa5HHPaXPTOnTub5gTAu6VNpUhJSdEvv/ximsdu+vTpatWqld180XXr1mnq1KkaP368gcn+MnnyZA0fPtx0b4TTpJ2Y/cMPP5juxOxr167Zlm708vLStWvX5OXlZaqrPo4fP16jRo3SpEmT1LVrV6Pj2ISGhmrr1q3q16+fNmzYoHHjxhkdyaZr165q27atTpw4YapPvu6c1GDWCQ5JSUk6evSoKleurKNHjxrySfpDWaqjoqK0bNkyWSwWXbt2TadOndKhQ4eMjmUzf/58oyNkqHjx4nriiSeMjnFP/v7+2rhxo90UBjO8WTLbPNF7MeMySXcy81W80k5+kqTAwEDTjKb/+uuv6dZp79ixo2kuciGlzs8MDAw0OkaG0k7Mtlgspjsxu0WLFurcubNq1aqlQ4cOqXnz5lq5cqX8/f2NjmYTHx+vW7du2a5eaBb16tVTpUqV5ObmJh8fH1Ot8PLbb78pIiJCMTExatOmjfz9/e1ONDbKnQXVTNM+08TGxmro0KEaPXq0Ll++rKJFi2rixIkuz/FQlurIyEj99NNPWrFihY4fP66goCCjI9lxd3c3fLJ9RgoVKqTw8HC7NW+Dg4MNTpXqyJEjOnLkiO22mZYOMzszLpN0pwIFCujXX3/Vr7/+attmllJ959Jhly9f1pUrVwxM85eMLjJkpmllhQsX1vjx4+3eCJvp9XjIkCF2J2aPGTPG6Eg2AwYMUIsWLXTixAl16NBBlSpVUlRUlDp37mx0NEmpI67Lli1T48aN1bRpU9WrV8/oSDahoaFq3LixfvzxRyUnJ2vz5s3617/+ZXQsSdLEiRM1ZcoUhYWFKSgoSL179zZFqT58+LA6deqklJQU/f7777avzfBJ+ooVK7R48WJ5eHgoLCzM0DfqD1WpTkhIsF0gxNPTU7Gxsdq6davpLhsdFhZmN9l+zJgxpllLO20NTbMUhzvdfaLJvS5Zinsz4zJJdzLDHOWM3Dmi7+XlpcmTJxuY5i9+fn46dOiQatasadt26NAh5c+f38BU9ooVKyZJOnv2rCTzjYCdP39eX3/9tWlPzP7uu+8UHx+vEydO6N///rcpTsxOU6JECbVs2VJS6snav/zyi8GJ/nL27Fm1bdtWH330ke2ERTMpU6aMLBaLChYsKB8fH6PjSJI+//xzoyNk6IsvvtBXX32l2NhYhYaGUqpdpXnz5nr++ec1bdo0lS1bVr179zZdoZbMMdn+bmlLNbVu3droKBky+5UyzezOZZLeeecdUyyTdKc7R6Wjo6NVunRpffnllwYm+svAgQPTLW1mtVrl6elpYKrU0bh+/fqpUaNGKl26tM6cOaOdO3dq3rx5hua6U0hIiKKiomS1WiVJFy9eNDiRvbVr1+qFF14wzbzWO5n1xOwffvhBv//+u5YuXaqXX35ZUuo0mpUrV+qLL74wOF0qq9WqTZs2qWLFioqKirKtlW4G+fPn1+rVq3Xr1i1t3LhRvr6+RkeS9NeydWbk5eUlLy8vFSxY0PZaYpSHqlT36NFDX3zxhc6ePaugoCAm29+HJUuWaNSoUenmaJppioXZr5RpZuPHj9eFCxfUqlUrradrLUAAABXjSURBVF+/Xu+++67RkezcOcXi7NmzmjNnjoFp7PXt21cXL15U+fLldfLkSeXOnVuJiYkaPny42rZta1iuUqVK6aOPPtK3336rP//8U7Vq1dKQIUOUJ08ewzLdbezYsdqzZ49u376t27dvq3Tp0qaa852QkKB27drZzeU3y1KOZj0xO3/+/Lpy5YqsVqvtQjVubm4aPny4wcn+0rt3b23cuFGjRo1SZGSkQkJCjI5kM3nyZM2fP18FChTQzz//rEmTJhkdKUcxutc9lOtU79mzR+vWrdO2bdsUFBSktm3bqlKlSkbHso0GHzlyRGFhYbp06ZKKFSumNm3amO7jKTPq1auXFi1apNDQUE2dOlVdu3bVhx9+aHSsHOHq1auaN2+eTp06JX9/f/Xt29dU0wTuFhwcbHfpbSP169dPkyZNUsGCBXX9+nWFhYVpwoQJevXVV7Vu3Tqj45lap06dtGrVKkVERCgkJESvv/66KdYLTluTd8+ePbp48aJtmopkjkuBS6nlq3bt2qY7MTvtHJvo6Gj5+fnZtpth7m1OEB4ern/84x8KCAgw1fkPZvbEE08oICBAKSkp2rVrlwICAmzfc/Wb4IdqpDpNw4YN1bBhQ924cUOfffaZQkND9emnnxodS6+++qqWLVumqlWr6uOPP1ZKSormzZunJUuWmKZUz5kzRx9++KHdk/3OUUQjmf1KmWYWEhKiZ599VkFBQdq3b59CQ0O1YMECo2PZ3HkRmEuXLqlQoUIGJ/rL1atXbdMD0kbp/Pz8TLVShFn5+PjYVoYww0e3adLW5G3YsKF69Ohhmk/j7nTkyBH9+uuvtpG5hIQEU7zRNNMlozOSNp0sJSVF169fN9V0sv9r797jar7/OIC/ju5uozV7sIiMmjCWsThNReTSTpdHiZyQmNsD5aRa2ZqF3B4y9rC5tOjC6FHZjHZJUo1cZ2F6oCiTSrmULqdzzuf3R4++v1IdKdv3e/J+/nW+55zv5/v6lo73+Xw/389HIpHg5MmT2LlzJ0xMTDBp0iRBzS0vRBEREdxjvlf0fC2L6nrdu3eHVCoVzATmS5cu5Qrr2tpa+Pv7Q0dHBwkJCXxH46SmpiI1NVWQY9HDwsJQUFCAVatWITIyUlBzj2qCWbNmAQDMzc2RnJzMc5o6K1euRERERKMPSj09PUHN/T1kyBD4+flhxIgR+PPPP/Hee+/h+PHjgir8hcrc3BxRUVEwMjKCTCaDQqHgOxIAYc/JW/83ER0d3Wj1XaH8Pybksbf1hDyczNLSEv3794e5uTliY2Px5ZdfUlH9AkK5egS85kW10Dg4OECpVGLevHl4+vQpvLy8BDVhPlA3pV5LU3XxTVtbG1lZWcjLy8OgQYPwwQcf8B1JY5iamuLo0aP46KOPcO3aNfTo0YNbXZHPS8r1VxuE9KH5vNDQUKSkpOD27dv45JNPYGNjg9zcXEFMgyV0vr6+qK6uhp6eHk6dOoX333+f70gAhD0nb2lpKfc4LS2NK6qFllNTvPPOO8jNzeU7BkcikaBTp05wdHTE2rVrBTE0lbSeMKuj19i0adOgUChw5MgRuLm58R2HU3/5/eHDh3B2dsagQYO4D3Gh3Ljj5+cHU1NTWFtb49KlSwgKCsKWLVv4jqURcnNzkZubi8jISGhpaaFLly74/PPPeb8RtaCgoMXLyX5+fv9xmubVDx3r1asXHj9+jKSkpEarGJKWubm5YdCgQZg0aRLGjx8PXV1dviMBEPacvA0JrRddUzw/nMzIyIjnRP+3YMECZGRkIC0tDUVFRRCLxY0WmCLCRkW1gNT/oTPGkJ+fj1mzZnErPfFduM6YMQN5eXlwcXGBjo4Ozp8/D0NDQ5iamvKaq6HHjx9DJpMBqJuKsH44A2nZtWvXEBwcjMOHD+PUqVMIDQ1Ft27duIUl+Kavry+Im6/UuX37NoC6Aufvv/9Gjx49qKhupcTEROTk5ODkyZPYu3cvevXqha+//prvWIKek1fIvehCp1AocPLkSbi7u3P3POjq6mLfvn08J/u/6dOnY9KkScjKysLu3btx/PhxpKen8x2LtBIV1QLScNwo34Ptn3fu3DncvHkTGzduhIGBAfr06YPw8HCUlpZizJgxfMcDALz77ru4ePEiLC0tkZOTgz59+qC2thaMMcH0gAnNtm3bEB4eDl1dXURERGDPnj0wMTGBj4+PIIpqIyMjODs78x1DrVWrVnGPGWP49NNPeUyjWW7evIkzZ87g8uXLYIxxi0vxTcjjgm/duoVVq1Zxvej1j+u/3JGWyWQyaGlp4eHDh7C3t4exsTF8fX3h5eXFdzTOokWLcP/+fQwbNgwrVqzAqFGj+I5EXgIV1QIi5HGjp0+fxuHDh7meEWNjY2zbtg0eHh6CWcXr4sWLyMjIgI6ODjeLwOTJkyESiZCSksJzOmFijMHc3BxFRUWoqqqChYUFAOH0gAnphsSWNFy5s6SkBPfu3eMxjWZxd3eHiYkJfH19MX78eL7jaISWZjoQWkeMEOXn5yMhIQFyuRyurq7Q0dHBgQMHMHDgQL6jcVcN69cMCA0NxYULFxAQEAA7Ozu+45FWoqKatIqBgUGTQktHR0cwS6gCjZeLJq2jUqkAAOnp6dzcnnK5HJWVlXzG4gQEBPAd4YUcHBy4YVv6+vrw8fHhO5LGyMrKwoULF5Ceno7IyEj06tULmzdv5juWoAm580XounbtCqBuyIdKpUJkZGSjubT5VH/V0NzcHIGBgY2uGlJRrTmoqCatYmBggIKCAvTt25d7rqCgQDA9mgCQkpKCuLg4bsjH48eP8dNPP/EdS9CsrKzg4eGBBw8eYNeuXcjPz0doaCimTp3KdzSNERERgeHDh3Pb586d4zGNZqmqqsKjR49QWlqKqqoqQd0wRjq2N998UzAFNdDyVUOa716zUFFNWkUmk2HJkiWwsrJC3759cf/+fWRkZGDjxo18R+N88803WLNmDQ4dOoQxY8YgMzOT70iCt3DhQkyYMAGGhobo2bMn8vPzMXPmTNjb2/MdTfAuXLiAW7duISoqCvPmzQNQ1/MfGxuLY8eO8ZxOM3h5eWHixInw9vaGubk533FIB9fcePR6fE8G0NJVw2fPnvEZi7wkKqpJqwwaNAhxcXFISUlBcXExLCwssHTpUu5ymhD07NkTI0eOxKFDh+Di4iKoRXOErOF4wn79+qFfv348ptEc3bt3x8OHDyGXy1FSUgKgbiy6v78/z8mE7/r16xgyZAgCAwMBAI8ePcKZM2cAoNESw4S8SkJaee95dNWwY6CimrRat27dBD1VWP1UfwqFAunp6VyhQ8i/YfDgwdzCDEK5WVdTnD59GkOGDGnyxVckElFRTf41Qh6PTlcNOwYRo9njiYarX/mvrKwMN27cgKWlJcLDw+Hu7k7f8sm/zsvLC99//z20tLT4jqLxysrKYGhoyHcMQghpE+qpJhrv888/b7SdnJyMsrIyFBYW8pSIvE4ePXoEa2trGBsbQyQSCW7VPSHbuXMnYmNjoVAoUFVVhb59++LEiRN8xyKEkDahnmrSIalUKkilUsTGxvIdhXRw//zzT5PnhLx4iJC4ubkhJiYG4eHhkEqlWLdunaBWtyOEkJdBPdWkw1Eqlbh48SKqq6v5jkJeA1paWli/fj1u376N/v37IygoiO9IGqN79+7Q09NDRUUFTE1NUVVVxXckQghpM5oAkXQ4NTU1iImJQUhICN9RyGsgJCQEEokEBw8ehLOzM4KDg/mOpDHefvttJCQkQF9fHxERESgvL+c7EiGEtBkN/yCEkHaQSqWIjo7mtj09PWnY0QusXLkSERERUCqVuH//Pnr27In4+HhYWVnBzMyM73iEENImNPyDEELaQalUIicnB2ZmZsjJyRHUKqNCVVZWBqBu6Ez9Kq1z587lMREhhLQf9VQTQkgbVVRU4MaNG9iwYQNKSkrQq1cvhIWF0eqAL2BrawtHR8dmX/Pz8/uP0xBCyKtBPdWEENIGMTExiIyMhLa2NkJCQvDxxx/zHUlj6OvrY8CAAXzHIISQV4qKakIIaYNjx44hOTkZFRUVWL16NRXVL8HIyAjOzs58xyCEkFeKZv8ghJA20NXVha6uLgwNDVFbW8t3HI0ydOhQviMQQsgrR0U1IYS0E92a8nICAgL4jkAIIa8c3ahICCFtMHbsWFhZWYExhrNnz8LKyop7bevWrTwmI4QQwgcqqgkhpA3OnTvX4mujR4/+D5MQQggRAiqqCSGEEEIIaScaU00IIYQQQkg7UVFNCCGEEEJIO1FRTQgRnHv37mHkyJFq31NYWAhra2tuyevmVFRUYP78+aiurm5TjvLycnh5eXHbEokET58+bVNbDWVlZWH69OntbkdTJCQkwMbGBvPnz2/1Pt7e3mp/t83Jzs6GnZ3dy8ZrJCQkBFevXlX7nqioKCQlJbXrOISQjoeKakKIxklKSoKnpyeKi4vVvm/Lli1wc3ODvr5+m47z5MkTZGdnc9tHjx5F9+7d29TW6ywpKQm+vr7Yt29fq/fJzMz8FxO17I8//njhFIlSqRT79+9HSUnJf5SKEKIJqKgmhGiUoqIi/P777y8s0AoLC5GamoqJEycCAAIDAxEWFgapVAp7e3ssW7YMz549AwDEx8fDzc0NTk5OsLW1RVxcHAAgKCgI1dXVkEgkUCqVMDMzQ1lZGTw8PPDLL79wx9q8eTM2b94MADhy5AhcXFzg5OSEuXPn4vbt22pz5uXlYd68eXB3d4etrS0WL16MmpoaAMCwYcOwY8cOeHh4wM7OjsulVCqxYcMG2Nvbw8XFBaGhoZBKpQDqCr7k5GSu/YbbLZ2nuvbKy8sRGBgIFxcXODo6Yv369VAoFE3Oo7y8HDKZDNOnT4ejoyM2bdoEhUKB9evXIzs7G9u3b0dUVFSjfZ49e4bly5dDIpHA2dkZISEhUKlUCAoKAgDMmTMHhYWFsLOza/TlpuF2XFwcJk+eDFdXV+586u3atQvOzs6QSCRYsmQJioqKuJ/J1q1b4enpCTs7OwQHB0OlUmHbtm0oLi6GTCbDlStX8Ouvv8LZ2RkuLi5wc3PD+fPnAQBaWlqYMmUK9uzZo/Z3Swh5zTBCCBGYgoICNmLEiBe+b/Dgway0tLTZ16Kjo1lAQAC3HRAQwGbMmMFqamqYXC5nTk5OLD4+nlVUVDB3d3dWVlbGGGPs8uXL3LGfz1F/vPj4eLZw4ULGGGMKhYKJxWKWl5fHsrKy2KxZs1hlZSVjjLH09HTm4ODQJNvZs2fZtGnTGGOMhYeHs6SkJMYYY3K5nE2fPp0lJydzx4uOjmaMMZadnc2GDh3Kqqur2cGDB5mnpyerrq5mNTU1zNvbm82ePZsxxtjs2bPZiRMnuGPVb6s7T3XtBQYGsgMHDnDnKpPJ2O7du5uc0+rVq9lXX33FVCoV18Z3333XbKZ6iYmJzNvbm2s7ODiY3blzp9HPmjHGbG1t2V9//cXtV799/fp1ZmVlxYqLixljjK1Zs4bZ2tpyba9cuZLV1tYyxhg7dOgQ8/Hx4fIsX76cKZVKVl5ezsRiMTtz5kyTY02YMIFdvnyZ+13u2LGDy3D16lVmY2PT5JwIIa8vbb6LekII+Tfk5uaiX79+jZ6ztraGrq4uAGDw4MF48uQJunTpgm+//RZpaWm4c+cObty4gcrKSrVtT506FZs2bUJJSQmuX7+O/v37o3///jh8+DDu3r0LDw8P7r1Pnz7F48eP0aNHj2bb8vf3R2ZmJvbs2YM7d+6guLi40fEnTJgAALCwsIBcLkdlZSXS0tIgkUigp6cHAJgxYwaio6PVZlZ3nuraO3XqFLKzsxEfHw8ALY5PP336NA4ePAiRSARdXV14eHhg//79WLhwYYuZLC0tsW3bNkilUowdOxZz5syBiYmJ2vNo6MyZMxg3bhzeeustLndGRgYAIDU1FdnZ2XB1dQUAqFQqVFVVcfva2tqiU6dO6Nq1K0xMTPDkyZMm7U+bNg3Lli3D+PHjMW7cOCxYsIB7zdjYGPfv30dNTQ33cyOEvN6oqCaEdEgikQgqlarRcw3HVotEIjDG8ODBA8yYMQPu7u6wtLSEg4MDUlNT1bZtYGCAyZMn49ixY7h8+TLc3NwA1BVuEokE/v7+3HZxcTHeeOONFtvy8/ODUqnElClTYGNjg8LCwkZjeusLNpFIBKBuSXRt7cYf3Z06NR7J13D/2tpaAFB7nuraU6lU2L59OwYOHAig7ktCfZaGVCpVo+dVKlWzw0Qa6tu3L3777TdkZWXh7NmzmDdvHtauXdvszYYNz0kulzf7vJaWVqPj+/j4YNasWdw+DQvn5v4tPM/X1xeurq7IzMxEQkICIiMjuS8XOjo6EIlEzf4sCCGvJxpTTQjpkAYMGICCgoIXvu/q1aswNDTEkiVLIBaLuUJTqVRCW1sbSqWy2YLL3d0diYmJuHTpEiZPngwAEIvF+Pnnn7kbKA8ePIg5c+aoPX5GRgaWLl2KqVOnAgCuXLkCpVKpdp/x48fjxx9/hFwuh0KhQGJiIveaoaEhN3vFrVu3kJOT88LzVNeeWCxGVFQUGGOQy+VYvHgxYmJimmQSi8WIiYnh3nf48GGMHTtW7XnExcUhKCgIYrEY/v7+EIvFuH79OoC6Arm+KG94TllZWdwNguPGjUNmZiYePHgAAE1yx8fHo6KiAgCwfft2rF69Wm2ehsdVKBSws7NDVVUVZs6ciS+++AI5OTlcQV9QUABjY2PuygchhFBPNSFEkCorK5tMq3fo0CGYmZm1av+JEydi7969UCqVjXownzdu3DjEx8fDwcEBIpEIo0ePhqGhIe7evQsTExMMHz4c06ZNQ2xsbKP9hg4dCi0tLTg4OHC9yWKxGAsWLIC3tzdEIhG6du2KnTt3qu3N9PX1xdKlS9G5c2d07doVH374IfLz89Wem4uLC/Ly8uDk5ITOnTvD2NgYBgYGAIDFixcjMDAQaWlpMDU1xahRo154nuraCw4Oxrp16+Do6Ija2lqMHTsWPj4+TTKFhIQgLCyMe5+1tTUWLVqk9jycnJxw7tw5TJ06FQYGBujduzd3g6SDgwOkUil27NgBmUyG0NBQ/PDDD7CwsICFhQUAwMzMDP7+/pgzZw66dOmC4cOHc227ubmhqKgI7u7uEIlE6N27N8LDw9XmAQB7e3v4+/sjNDQUn332GWQyGbS1tSESibB+/XquiE5PT4eDg8ML2yOEvD5omXJCSIe1Zs0aWFlZcb3AHUVGRgZKS0shkUgAAGFhYdDT0+OGnfDdXkenVCrh7OyMyMhIGBkZ8R2HECIQVFQTQjqsp0+fYsWKFdi1a1eb56oWoqKiIgQGBuLhw4dQqVQwNzdHaGgounXrJoj2OrqoqCh069aNuwmSEEIAKqoJIYQQQghpN7pRkRBCCCGEkHaiopoQQgghhJB2oqKaEEIIIYSQdqKimhBCCCGEkHaiopoQQgghhJB2+h/vnZWZ0t3E1AAAAABJRU5ErkJggg==\n",
      "text/plain": [
       "<Figure size 864x360 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "fig = plt.figure()\n",
    "df_native.plot(kind='bar', figsize=(12,5))   # Not saving output as a plot object\n",
    "plt.title('Figure 3: number of text by L1', fontsize=16)          # Must use plt.title(), plt.xlabel() etc. \n",
    "plt.xlabel('L1 (native language of students)', fontsize=12)\n",
    "plt.ylabel('Number of text', fontsize=12)\n",
    "#plt.show()\n",
    "\n",
    "fig.savefig('./image_files/part1_plot3.png')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Summary\n",
    "- From the chart above, we can find that most texts are written by Arabic, Korean and Chinese students.\n",
    "- The number of text by L1 are not distributed well."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Preprocessing for Linguistic Analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 35,
   "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>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "      <th>tokens</th>\n",
       "      <th>token_count</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>43439</th>\n",
       "      <td>gu2</td>\n",
       "      <td>5796</td>\n",
       "      <td>They see the online shopping do pollution but ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...</td>\n",
       "      <td>[They, see, the, online, shopping, do, polluti...</td>\n",
       "      <td>109</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>753</th>\n",
       "      <td>cq4</td>\n",
       "      <td>37</td>\n",
       "      <td>1-Mismanagement\\nSentence: So if they were nom...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>For each of the following words, write the sen...</td>\n",
       "      <td>[1-Mismanagement, Sentence, :, So, if, they, w...</td>\n",
       "      <td>198</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45003</th>\n",
       "      <td>ay8</td>\n",
       "      <td>5990</td>\n",
       "      <td>I was working out in the gym.GHABBAN.wav.mp3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What were  you doing at 8:00 this morning?</td>\n",
       "      <td>[I, was, working, out, in, the, gym.GHABBAN.wa...</td>\n",
       "      <td>7</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16101</th>\n",
       "      <td>ff4</td>\n",
       "      <td>2358</td>\n",
       "      <td>it's getting out of control. Something you sta...</td>\n",
       "      <td>Male</td>\n",
       "      <td>French</td>\n",
       "      <td>What does the phrase \"it's getting out of hand...</td>\n",
       "      <td>[it, 's, getting, out, of, control, ., Somethi...</td>\n",
       "      <td>17</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23749</th>\n",
       "      <td>cp5</td>\n",
       "      <td>3229</td>\n",
       "      <td>This tree has a very rigid trunk, I can not cu...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Write an original creative sentence using a vo...</td>\n",
       "      <td>[This, tree, has, a, very, rigid, trunk, ,, I,...</td>\n",
       "      <td>15</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      student_id  question_id  \\\n",
       "43439        gu2         5796   \n",
       "753          cq4           37   \n",
       "45003        ay8         5990   \n",
       "16101        ff4         2358   \n",
       "23749        cp5         3229   \n",
       "\n",
       "                                                    text  gender  \\\n",
       "43439  They see the online shopping do pollution but ...     NaN   \n",
       "753    1-Mismanagement\\nSentence: So if they were nom...  Female   \n",
       "45003       I was working out in the gym.GHABBAN.wav.mp3     NaN   \n",
       "16101  it's getting out of control. Something you sta...    Male   \n",
       "23749  This tree has a very rigid trunk, I can not cu...    Male   \n",
       "\n",
       "      native_language                                           question  \\\n",
       "43439             NaN  Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...   \n",
       "753            Arabic  For each of the following words, write the sen...   \n",
       "45003             NaN         What were  you doing at 8:00 this morning?   \n",
       "16101          French  What does the phrase \"it's getting out of hand...   \n",
       "23749          Arabic  Write an original creative sentence using a vo...   \n",
       "\n",
       "                                                  tokens  token_count  \n",
       "43439  [They, see, the, online, shopping, do, polluti...          109  \n",
       "753    [1-Mismanagement, Sentence, :, So, if, they, w...          198  \n",
       "45003  [I, was, working, out, in, the, gym.GHABBAN.wa...            7  \n",
       "16101  [it, 's, getting, out, of, control, ., Somethi...           17  \n",
       "23749  [This, tree, has, a, very, rigid, trunk, ,, I,...           15  "
      ]
     },
     "execution_count": 35,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.head()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "- Type token ratio analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 36,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_eli_short['types'] = df_eli_short['tokens'].map(set)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 37,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_eli_short['type_count'] = df_eli_short['types'].map(len)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 38,
   "metadata": {},
   "outputs": [],
   "source": [
    "df_eli_short['TTR'] = df_eli_short['type_count'] / df_eli_short['token_count']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 39,
   "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>student_id</th>\n",
       "      <th>question_id</th>\n",
       "      <th>text</th>\n",
       "      <th>gender</th>\n",
       "      <th>native_language</th>\n",
       "      <th>question</th>\n",
       "      <th>tokens</th>\n",
       "      <th>token_count</th>\n",
       "      <th>types</th>\n",
       "      <th>type_count</th>\n",
       "      <th>TTR</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>43439</th>\n",
       "      <td>gu2</td>\n",
       "      <td>5796</td>\n",
       "      <td>They see the online shopping do pollution but ...</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...</td>\n",
       "      <td>[They, see, the, online, shopping, do, polluti...</td>\n",
       "      <td>109</td>\n",
       "      <td>{goods, a, see, as, with, efficient, those, th...</td>\n",
       "      <td>79</td>\n",
       "      <td>0.724771</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>753</th>\n",
       "      <td>cq4</td>\n",
       "      <td>37</td>\n",
       "      <td>1-Mismanagement\\nSentence: So if they were nom...</td>\n",
       "      <td>Female</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>For each of the following words, write the sen...</td>\n",
       "      <td>[1-Mismanagement, Sentence, :, So, if, they, w...</td>\n",
       "      <td>198</td>\n",
       "      <td>{An, construction, proves, 4, poor, .that, wou...</td>\n",
       "      <td>116</td>\n",
       "      <td>0.585859</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>45003</th>\n",
       "      <td>ay8</td>\n",
       "      <td>5990</td>\n",
       "      <td>I was working out in the gym.GHABBAN.wav.mp3</td>\n",
       "      <td>NaN</td>\n",
       "      <td>NaN</td>\n",
       "      <td>What were  you doing at 8:00 this morning?</td>\n",
       "      <td>[I, was, working, out, in, the, gym.GHABBAN.wa...</td>\n",
       "      <td>7</td>\n",
       "      <td>{in, gym.GHABBAN.wav.mp3, out, was, I, working...</td>\n",
       "      <td>7</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16101</th>\n",
       "      <td>ff4</td>\n",
       "      <td>2358</td>\n",
       "      <td>it's getting out of control. Something you sta...</td>\n",
       "      <td>Male</td>\n",
       "      <td>French</td>\n",
       "      <td>What does the phrase \"it's getting out of hand...</td>\n",
       "      <td>[it, 's, getting, out, of, control, ., Somethi...</td>\n",
       "      <td>17</td>\n",
       "      <td>{you, n't, ., Something, getting, start, and, ...</td>\n",
       "      <td>15</td>\n",
       "      <td>0.882353</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23749</th>\n",
       "      <td>cp5</td>\n",
       "      <td>3229</td>\n",
       "      <td>This tree has a very rigid trunk, I can not cu...</td>\n",
       "      <td>Male</td>\n",
       "      <td>Arabic</td>\n",
       "      <td>Write an original creative sentence using a vo...</td>\n",
       "      <td>[This, tree, has, a, very, rigid, trunk, ,, I,...</td>\n",
       "      <td>15</td>\n",
       "      <td>{a, has, ., tree, can, ,, very, trunk, not, by...</td>\n",
       "      <td>15</td>\n",
       "      <td>1.000000</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "      student_id  question_id  \\\n",
       "43439        gu2         5796   \n",
       "753          cq4           37   \n",
       "45003        ay8         5990   \n",
       "16101        ff4         2358   \n",
       "23749        cp5         3229   \n",
       "\n",
       "                                                    text  gender  \\\n",
       "43439  They see the online shopping do pollution but ...     NaN   \n",
       "753    1-Mismanagement\\nSentence: So if they were nom...  Female   \n",
       "45003       I was working out in the gym.GHABBAN.wav.mp3     NaN   \n",
       "16101  it's getting out of control. Something you sta...    Male   \n",
       "23749  This tree has a very rigid trunk, I can not cu...    Male   \n",
       "\n",
       "      native_language                                           question  \\\n",
       "43439             NaN  Chapter 5: \"A Cleaner Way to Shop?\"  \\n\\nChoos...   \n",
       "753            Arabic  For each of the following words, write the sen...   \n",
       "45003             NaN         What were  you doing at 8:00 this morning?   \n",
       "16101          French  What does the phrase \"it's getting out of hand...   \n",
       "23749          Arabic  Write an original creative sentence using a vo...   \n",
       "\n",
       "                                                  tokens  token_count  \\\n",
       "43439  [They, see, the, online, shopping, do, polluti...          109   \n",
       "753    [1-Mismanagement, Sentence, :, So, if, they, w...          198   \n",
       "45003  [I, was, working, out, in, the, gym.GHABBAN.wa...            7   \n",
       "16101  [it, 's, getting, out, of, control, ., Somethi...           17   \n",
       "23749  [This, tree, has, a, very, rigid, trunk, ,, I,...           15   \n",
       "\n",
       "                                                   types  type_count       TTR  \n",
       "43439  {goods, a, see, as, with, efficient, those, th...          79  0.724771  \n",
       "753    {An, construction, proves, 4, poor, .that, wou...         116  0.585859  \n",
       "45003  {in, gym.GHABBAN.wav.mp3, out, was, I, working...           7  1.000000  \n",
       "16101  {you, n't, ., Something, getting, start, and, ...          15  0.882353  \n",
       "23749  {a, has, ., tree, can, ,, very, trunk, not, by...          15  1.000000  "
      ]
     },
     "execution_count": 39,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 32,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "cx0    5\n",
       "es9    5\n",
       "cp5    4\n",
       "gv1    4\n",
       "bs1    4\n",
       "cb8    3\n",
       "ac5    3\n",
       "fs4    3\n",
       "bq0    3\n",
       "dq9    3\n",
       "Name: student_id, dtype: int64"
      ]
     },
     "execution_count": 32,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.student_id.value_counts().head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 33,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    353.000000\n",
       "mean       1.393768\n",
       "std        0.715626\n",
       "min        1.000000\n",
       "25%        1.000000\n",
       "50%        1.000000\n",
       "75%        2.000000\n",
       "max        5.000000\n",
       "Name: student_id, dtype: float64"
      ]
     },
     "execution_count": 33,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.student_id.value_counts().describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 34,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count     492\n",
       "unique    353\n",
       "top       cx0\n",
       "freq        5\n",
       "Name: student_id, dtype: object"
      ]
     },
     "execution_count": 34,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.student_id.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 42,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count     492.000000\n",
       "mean      100.709350\n",
       "std       173.646359\n",
       "min         1.000000\n",
       "25%         8.000000\n",
       "50%        24.500000\n",
       "75%       124.250000\n",
       "max      1262.000000\n",
       "Name: token_count, dtype: float64"
      ]
     },
     "execution_count": 42,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.token_count.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 41,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "count    492.000000\n",
       "mean       0.783395\n",
       "std        0.212235\n",
       "min        0.298103\n",
       "25%        0.597561\n",
       "50%        0.854497\n",
       "75%        1.000000\n",
       "max        1.000000\n",
       "Name: TTR, dtype: float64"
      ]
     },
     "execution_count": 41,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.TTR.describe()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 45,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "1.000000    0.333333\n",
       "0.857143    0.014228\n",
       "0.937500    0.012195\n",
       "0.900000    0.010163\n",
       "0.750000    0.010163\n",
       "Name: TTR, dtype: float64"
      ]
     },
     "execution_count": 45,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "df_eli_short.TTR.value_counts(normalize=True).head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 40,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pickle\n",
    "\n",
    "f = open('df_eli.pkl', 'wb')   # 'wb' instead 'w' for binary file\n",
    "pickle.dump(df_eli_short, f, -1)       # -1 specifies highest binary protocol\n",
    "f.close()  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Conclusion\n",
    "\n",
    "- There are 492 texts in the dataframe.\n",
    "- These texts are written by 353 students.\n",
    "- The max number of essays students wrote is 5.\n",
    "- Students write 1.39 essays in average.\n",
    "- The length of texts are not distributed well. Most of them (about 65%) are under 65 tokens. Half of them are under 24.5 tokens.\n",
    "- Most texts are written by Arabic, Korean and Chinese students.\n",
    "- TTR has a lot of bias. About 33% texts' TTR are 1. That is because there are only one word in the texts."
   ]
  },
  {
   "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": "3.6.8"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}