visualization.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "# Toxic Comment Visualization"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 48,
   "metadata": {},
   "outputs": [],
   "source": [
    "# load required libraries:\n",
    "# Remember to pip install spacy and scattertext \n",
    "# If errors still exist, open your terminal and type python -m spacy download en\n",
    "\n",
    "import joblib\n",
    "import nltk\n",
    "import spacy\n",
    "nlp = spacy.load('en_core_web_sm')\n",
    "\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import sklearn\n",
    "\n",
    "import vocabulary\n",
    "import tokenizer\n",
    "import vector\n",
    "import dataclean\n",
    "import lemmatizer\n",
    "import pos_tagger\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import matplotlib.gridspec as gridspec\n",
    "import seaborn as sns\n",
    "import scattertext as st\n",
    "from IPython.display import IFrame"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 49,
   "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>id</th>\n",
       "      <th>comment_text</th>\n",
       "      <th>toxic</th>\n",
       "      <th>severe_toxic</th>\n",
       "      <th>obscene</th>\n",
       "      <th>threat</th>\n",
       "      <th>insult</th>\n",
       "      <th>identity_hate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0000997932d777bf</td>\n",
       "      <td>Explanation\\nWhy the edits made under my usern...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>000103f0d9cfb60f</td>\n",
       "      <td>D'aww! He matches this background colour I'm s...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>000113f07ec002fd</td>\n",
       "      <td>Hey man, I'm really not trying to edit war. It...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0001b41b1c6bb37e</td>\n",
       "      <td>\"\\nMore\\nI can't make any real suggestions on ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0001d958c54c6e35</td>\n",
       "      <td>You, sir, are my hero. Any chance you remember...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 id                                       comment_text  toxic  \\\n",
       "0  0000997932d777bf  Explanation\\nWhy the edits made under my usern...      0   \n",
       "1  000103f0d9cfb60f  D'aww! He matches this background colour I'm s...      0   \n",
       "2  000113f07ec002fd  Hey man, I'm really not trying to edit war. It...      0   \n",
       "3  0001b41b1c6bb37e  \"\\nMore\\nI can't make any real suggestions on ...      0   \n",
       "4  0001d958c54c6e35  You, sir, are my hero. Any chance you remember...      0   \n",
       "\n",
       "   severe_toxic  obscene  threat  insult  identity_hate  \n",
       "0             0        0       0       0              0  \n",
       "1             0        0       0       0              0  \n",
       "2             0        0       0       0              0  \n",
       "3             0        0       0       0              0  \n",
       "4             0        0       0       0              0  "
      ]
     },
     "execution_count": 49,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# load the comment data\n",
    "train = pd.read_csv('train.csv')\n",
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "x=train.iloc[:,2:].sum()\n",
    "\n",
    "plt.figure(figsize=(8,4))\n",
    "ax= sns.barplot(x.index, x.values, alpha=0.8)\n",
    "plt.ylabel('number for each type', fontsize=12)\n",
    "plt.xlabel('Type ', fontsize=12)\n",
    "\n",
    "rects = ax.patches\n",
    "labels = x.values\n",
    "for rect, label in zip(rects, labels):\n",
    "    height = rect.get_height()\n",
    "    ax.text(rect.get_x() + rect.get_width()/2, height + 5, label, ha='center', va='bottom')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "rowsums=train.iloc[:,2:].sum(axis=1)\n",
    "x=rowsums.value_counts()\n",
    "\n",
    "#plot\n",
    "plt.figure(figsize=(8,4))\n",
    "ax = sns.barplot(x.index, x.values, alpha=0.8)\n",
    "plt.ylabel('number of comments', fontsize=12)\n",
    "plt.xlabel('number of tags ', fontsize=12)\n",
    "\n",
    "#adding the text labels\n",
    "rects = ax.patches\n",
    "labels = x.values\n",
    "for rect, label in zip(rects, labels):\n",
    "    height = rect.get_height()\n",
    "    ax.text(rect.get_x() + rect.get_width()/2, height + 5, label, ha='center', va='bottom')\n",
    "\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"Explanation\\nWhy the edits made under my username Hardcore Metallica Fan were reverted? They weren't vandalisms, just closure on some GAs after I voted at New York Dolls FAC. And please don't remove the template from the talk page since I'm retired now.89.205.38.27\""
      ]
     },
     "execution_count": 17,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.comment_text[0]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "\"D'aww! He matches this background colour I'm seemingly stuck with. Thanks.  (talk) 21:51, January 11, 2016 (UTC)\""
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "train.comment_text[1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 50,
   "metadata": {},
   "outputs": [],
   "source": [
    "# shuffle and take the first 10k rows for the ease of preprocessing\n",
    "train = train.sample(frac=1).reset_index(drop=True)\n",
    "train = train[:10000]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 51,
   "metadata": {
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "# remove punctuation\n",
    "cleaner = dataclean.DataClean()\n",
    "train.comment_text = train.comment_text.apply(cleaner.clean)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 52,
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "# NLTK stoplist with 3136 words\n",
    "STOPLIST = set(nltk.corpus.stopwords.words())\n",
    "tkn = tokenizer.Tokenizer()\n",
    "\n",
    "\n",
    "def remove_stopwords(text):\n",
    "    # remove stop words from text\n",
    "    tokens = tkn.tokenize(text)\n",
    "    words = [word for word in tokens if word not in STOPLIST]\n",
    "    return \" \".join(words)\n",
    "\n",
    "\n",
    "# remove stopwords from comments\n",
    "train.comment_text = train.comment_text.apply(remove_stopwords)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 53,
   "metadata": {
    "scrolled": true
   },
   "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>id</th>\n",
       "      <th>comment_text</th>\n",
       "      <th>toxic</th>\n",
       "      <th>severe_toxic</th>\n",
       "      <th>obscene</th>\n",
       "      <th>threat</th>\n",
       "      <th>insult</th>\n",
       "      <th>identity_hate</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>f297eb61033cfb05</td>\n",
       "      <td>(references, really, canonical, ?, reference, ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>5688e3c95d4b6c88</td>\n",
       "      <td>(exactly, ., clear, guidelines, place, self, p...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>2b800bc2c4d5eb5e</td>\n",
       "      <td>(question, stats, minor, quibble, ., article, ...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>219ac63222da4ada</td>\n",
       "      <td>(links, sites, erroneous, claims, although, an...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>034380b8a8d44868</td>\n",
       "      <td>(commonwealth, instead, dominion, ?, based, up...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 id                                       comment_text  toxic  \\\n",
       "0  f297eb61033cfb05  (references, really, canonical, ?, reference, ...      0   \n",
       "1  5688e3c95d4b6c88  (exactly, ., clear, guidelines, place, self, p...      0   \n",
       "2  2b800bc2c4d5eb5e  (question, stats, minor, quibble, ., article, ...      0   \n",
       "3  219ac63222da4ada  (links, sites, erroneous, claims, although, an...      0   \n",
       "4  034380b8a8d44868  (commonwealth, instead, dominion, ?, based, up...      0   \n",
       "\n",
       "   severe_toxic  obscene  threat  insult  identity_hate  \n",
       "0             0        0       0       0              0  \n",
       "1             0        0       0       0              0  \n",
       "2             0        0       0       0              0  \n",
       "3             0        0       0       0              0  \n",
       "4             0        0       0       0              0  "
      ]
     },
     "execution_count": 53,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# parse the comments\n",
    "train.comment_text = train.comment_text.apply(nlp)\n",
    "\n",
    "train.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 54,
   "metadata": {},
   "outputs": [],
   "source": [
    "def create_corpus(dataframe,label):\n",
    "    \"\"\"\n",
    "    a function that convert dataframe into scattertext corpus\n",
    "    :param dataframe: dataframe to be converted\n",
    "    :param label: type of toxicity behavior\n",
    "    :return: a corpus that is required as an input of scattertext plot\n",
    "    \"\"\"\n",
    "    dataframe[label] = dataframe[label].replace(0, 'Not '+label)\n",
    "    dataframe[label] = dataframe[label].replace(1, label)\n",
    "    corpus = st.CorpusFromParsedDocuments(dataframe,\n",
    "                                          category_col=label,\n",
    "                                          parsed_col='comment_text').build()\n",
    "    return corpus\n",
    "\n",
    "\n",
    "def visualize_chart(corpus, toxicity, not_toxicity):\n",
    "    \"\"\"\n",
    "    a function that visualize the text using the scattertext library\n",
    "    :param corpus: data to be visualized\n",
    "    :param toxicity: type of toxicity behavior\n",
    "    :not_toxicity:\n",
    "    :output: a scattertext plot\n",
    "    \"\"\"\n",
    "    html = st.produce_scattertext_explorer(corpus,\n",
    "                                        category=toxicity,\n",
    "                                        category_name=toxicity,\n",
    "                                        not_category_name=not_toxicity,\n",
    "                                        width_in_pixels=1000,\n",
    "                                        minimum_term_frequency=5,\n",
    "                                        transform=st.Scalers.log_scale_standardize)\n",
    "    filename='visualization/'+toxicity+'_vs_'+not_toxicity+'.html'\n",
    "    with open(filename, 'wb') as f:\n",
    "        f.write(html.encode('utf-8'))\n",
    "        IFrame(src=filename, width = 1200, height=700)\n",
    "\n",
    "# plot the characteristic terms used by different comment group\n",
    "toxicity_list = ['toxic', 'severe_toxic', 'obscene', 'threat', 'insult', 'identity_hate']\n",
    "for toxicity in toxicity_list:\n",
    "    corpus = create_corpus(train, toxicity)\n",
    "    visualize_chart(corpus, toxicity, \"Not \"+toxicity)\n"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 1
}