Course-3/Week-3/C3_W3_Lab-3/C3_W3_Lab_3_Haiti_Earthquake_Design_2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Topic Modeling: Design Phase - Part 2"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In this notebook you will perform topic modeling by using a technique known as [Latent Dirichlet Allocation](https://towardsdatascience.com/latent-dirichlet-allocation-lda-9d1cd064ffa2) (LDA for short). Then, you will quantitatively compare the different models using a metric suited to measure the coherence of the topics generated.\n",
    "\n",
    "\n",
    "The steps you will complete in this lab are:\n",
    "1. Import Python packages\n",
    "2. Load the dataset\n",
    "3. Represent the data as Bag-of-Words\n",
    "4. Get a base LDA model\n",
    "5. Introducing the Coherence Score Metric\n",
    "6. Topic Modeling with LDA\n",
    "7. Select and evaluate the final LDA model\n",
    "8. Removing additional words (optional)\n",
    "\n",
    "**Note:** In all other case studies in this specialization, you have worked with examples of supervised learning, where you trained a model to learn the mapping from and input A to an output B. In this lab, you'll be developing an unsupervised learning model, which is to say, there are no labels or explicit right answers to train on. Your model will be looking for statistical similarities between the text of different messages in order to group them into topics. This sort of task is common in unsupervised learning projects, where you're often attempting to discover relationships or patterns in your data and draw inferences based on those patterns or relationships.  "
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 1. Import Python packages\n",
    "\n",
    "Run the next cell to import that Python packages you'll need for this lab.\n",
    "\n",
    "Note the `import utils` line. This line imports the functions that were specifically written for this lab. If you want to look at what these functions are, go to `File -> Open...` and open the `utils.py` file to have a look."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "[nltk_data] Downloading package averaged_perceptron_tagger to\n",
      "[nltk_data]     /home/jovyan/nltk_data...\n",
      "[nltk_data]   Unzipping taggers/averaged_perceptron_tagger.zip.\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "All packages imported successfully!\n"
     ]
    }
   ],
   "source": [
    "import warnings\n",
    "warnings.filterwarnings('ignore')\n",
    "import nltk\n",
    "nltk.download('averaged_perceptron_tagger')\n",
    "from nltk.tokenize import word_tokenize\n",
    "from nltk.stem import WordNetLemmatizer\n",
    "from termcolor import colored\n",
    "import random\n",
    "import gensim\n",
    "import pandas as pd\n",
    "\n",
    "import utils\n",
    "\n",
    "\n",
    "print('All packages imported successfully!')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 2. Load the dataset\n",
    "\n",
    "Begin by loading in the cleaned dataframe from the last ungraded lab (that is saved in a `.pkl` file):"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "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>split</th>\n",
       "      <th>message</th>\n",
       "      <th>original</th>\n",
       "      <th>genre</th>\n",
       "      <th>related</th>\n",
       "      <th>PII</th>\n",
       "      <th>request</th>\n",
       "      <th>offer</th>\n",
       "      <th>aid_related</th>\n",
       "      <th>...</th>\n",
       "      <th>earthquake</th>\n",
       "      <th>cold</th>\n",
       "      <th>other_weather</th>\n",
       "      <th>direct_report</th>\n",
       "      <th>event</th>\n",
       "      <th>actionable_haiti</th>\n",
       "      <th>date_haiti</th>\n",
       "      <th>message_tokens</th>\n",
       "      <th>num_words</th>\n",
       "      <th>num_tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>train</td>\n",
       "      <td>Weather update - a cold front from Cuba that c...</td>\n",
       "      <td>Un front froid se retrouve sur Cuba ce matin. ...</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-02-09</td>\n",
       "      <td>[weather, update, cold, front, cuba, could, pa...</td>\n",
       "      <td>13</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7</td>\n",
       "      <td>train</td>\n",
       "      <td>Is the Hurricane over or is it not over</td>\n",
       "      <td>Cyclone nan fini osinon li pa fini</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[hurricane]</td>\n",
       "      <td>9</td>\n",
       "      <td>1</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>train</td>\n",
       "      <td>says: west side of Haiti, rest of the country ...</td>\n",
       "      <td>facade ouest d Haiti et le reste du pays aujou...</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[say, west, side, haiti, rest, country, today,...</td>\n",
       "      <td>12</td>\n",
       "      <td>8</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>14</td>\n",
       "      <td>train</td>\n",
       "      <td>Information about the National Palace-</td>\n",
       "      <td>Informtion au nivaux palais nationl</td>\n",
       "      <td>direct</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[information, national, palace-]</td>\n",
       "      <td>5</td>\n",
       "      <td>3</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>train</td>\n",
       "      <td>Storm at sacred heart of jesus</td>\n",
       "      <td>Cyclone Coeur sacr de jesus</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[storm, sacred, heart, jesus]</td>\n",
       "      <td>6</td>\n",
       "      <td>4</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 48 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   id  split                                            message  \\\n",
       "0   2  train  Weather update - a cold front from Cuba that c...   \n",
       "1   7  train            Is the Hurricane over or is it not over   \n",
       "2  12  train  says: west side of Haiti, rest of the country ...   \n",
       "3  14  train             Information about the National Palace-   \n",
       "4  15  train                     Storm at sacred heart of jesus   \n",
       "\n",
       "                                            original   genre  related  PII  \\\n",
       "0  Un front froid se retrouve sur Cuba ce matin. ...  direct        1    0   \n",
       "1                 Cyclone nan fini osinon li pa fini  direct        1    0   \n",
       "2  facade ouest d Haiti et le reste du pays aujou...  direct        1    0   \n",
       "3                Informtion au nivaux palais nationl  direct        0    0   \n",
       "4                        Cyclone Coeur sacr de jesus  direct        1    0   \n",
       "\n",
       "   request  offer  aid_related  ...  earthquake  cold  other_weather  \\\n",
       "0        0      0            0  ...           0     0              0   \n",
       "1        0      0            1  ...           0     0              0   \n",
       "2        0      0            0  ...           0     0              0   \n",
       "3        0      0            0  ...           0     0              0   \n",
       "4        0      0            0  ...           0     0              0   \n",
       "\n",
       "   direct_report             event  actionable_haiti  date_haiti  \\\n",
       "0              0  haiti_earthquake                 0  2010-02-09   \n",
       "1              0  haiti_earthquake                 0  2010-01-17   \n",
       "2              0  haiti_earthquake                 0  2010-01-17   \n",
       "3              0  haiti_earthquake                 0  2010-01-17   \n",
       "4              0  haiti_earthquake                 0  2010-01-17   \n",
       "\n",
       "                                      message_tokens  num_words  num_tokens  \n",
       "0  [weather, update, cold, front, cuba, could, pa...         13           8  \n",
       "1                                        [hurricane]          9           1  \n",
       "2  [say, west, side, haiti, rest, country, today,...         12           8  \n",
       "3                   [information, national, palace-]          5           3  \n",
       "4                      [storm, sacred, heart, jesus]          6           4  \n",
       "\n",
       "[5 rows x 48 columns]"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "haiti_df = pd.read_pickle(\"./data/haiti_df_processed.pkl\")\n",
    "\n",
    "haiti_df.head(5)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 3. Represent the data as Bag-of-Words\n",
    "\n",
    "Now you will convert the messages in the dataset into their BOW representation and generate other useful information such as a list containing all words in the corpus:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Dictionary contains a total of 1604 unique words\n"
     ]
    }
   ],
   "source": [
    "# All of the messages in the dataset\n",
    "corpus = haiti_df.message_tokens\n",
    "\n",
    "# Create the dictionary\n",
    "corpus_dictionary = gensim.corpora.Dictionary(corpus)\n",
    "\n",
    "# Filter out words that occur less than 5 times overall\n",
    "corpus_dictionary.filter_extremes(no_below=5)\n",
    "\n",
    "# Create a bag of words out of the full corpus\n",
    "corpus_bow = [corpus_dictionary.doc2bow(doc) for doc in corpus]\n",
    "\n",
    "# All words in the dictionary\n",
    "all_words = list(corpus_dictionary.values())\n",
    "\n",
    "print(f\"Dictionary contains a total of {len(all_words)} unique words\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 4. Get a base LDA model\n",
    "\n",
    "There are several algorithms to perform topic modeling, as well as several Python libraries for this purpose. In this section you will perform topic modeling using Latent Dirichlet Allocation (LDA) with Gensim and pyLDAvis (for visualization). The three methods and libraries you will use are:\n",
    "\n",
    "Latent Dirichlet Allocation (LDA) is a generative probabilistic topic model that aims to uncover latent (or hidden) topics from a corpus. The corpus in this case is the collection of processed and tokenized messages. After topic modeling, each document is represented by a distribution over the topics that were found in the entire corpus.\n",
    "\n",
    "To interact with the LDA models you will be using the amazing library [Gensim](https://github.com/RaRe-Technologies/gensim). This library provides a lot of features for performing various NLP-related tasks. In fact, you already used it at the beginning of the lab to generate the BOW representation of the corpus.\n",
    "\n",
    "To train a LDA model on a corpus you can use the [LdaModel](https://radimrehurek.com/gensim/models/ldamodel.html) class. In the next cell you will use it to train the model (it is abstracted in the `utils.train_lda_model` fumction) on the corpus data."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_topics = 2\n",
    "lda_2_topics = utils.train_lda_model(num_topics=num_topics, corpus_bow=corpus_bow, corpus=corpus, corpus_dictionary=corpus_dictionary)\n",
    "utils.plot_top_words_lda(num_topics, lda_2_topics.model)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 5 Introducing the Coherence Score Metric\n",
    "\n",
    "A common metric to compare models that aim at capturing topics from a corpus is the `coherence score`. This metric arises from the idea that \"a set of statements or facts are said to be coherent if they support each other\". This quote is taken from the paper [Exploring the Space of Topic Coherence Measures](https://svn.aksw.org/papers/2015/WSDM_Topic_Evaluation/public.pdf), in case you want to look into the details on this subject.\n",
    "\n",
    "A nice and simple explanation can be also found here:\n",
    "\n",
    "- [Understanding Topic Coherence Measures](https://towardsdatascience.com/understanding-topic-coherence-measures-4aa41339634c)\n",
    "\n",
    "More resources:\n",
    "\n",
    "- [Evaluating topic coherence measures](https://mimno.infosci.cornell.edu/nips2013ws/nips2013tm_submission_7.pdf)\n",
    "\n",
    "To compute this metric the `compute_coherence_score` function is used. This function uses Gensim's [CoherenceModel](https://radimrehurek.com/gensim/models/coherencemodel.html) class to compute the score for any given list of topics and corpus.\n",
    "\n",
    "**This metric is defined in such a way that values closer to 0 reflect a greater coherence**. \n",
    "\n",
    "To test it you will use two examples. One has the actual 10 top words for the topics (taken from the model) and the other is a random collection of words from the corpus. Notice that both examples have two lists of words which means that they simulate having two topics:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Top words:\n",
      "    ['like', 'would', 'know', 'information', 'message', 'please', 'earthquake', 'want', 'give', 'thank']\n",
      "    ['help', 'need', 'food', 'please', 'people', 'u', 'water', 'find', 'tent', 'house']\n",
      "Coherence score for the text using calculated topics is: -2.347\n",
      "\n",
      "Random words:\n",
      "    ['recipient', 'club', 'bottle', '100g', 'kafou', 'peace', 'natural', 'stick', 'whith', 'gon']\n",
      "    ['sen', 'certificate', 'know', 'passport', 'creole', 'hand', 'day', 'unknown', 'drive', 'wait']\n",
      "Coherence score for the text using random topics is: -18.889\n"
     ]
    }
   ],
   "source": [
    "real_top_words = utils.get_top_words_lda(num_topics, lda_2_topics.model, num_words=10)\n",
    "\n",
    "random_topics = [\n",
    "    random.sample(all_words, 10),\n",
    "    random.sample(all_words, 10),    \n",
    "]\n",
    "\n",
    "print(f\"Top words:\")\n",
    "for word_list in real_top_words:\n",
    "    print(f\"    {word_list}\")\n",
    "print(f\"Coherence score for the text using calculated topics is: {utils.compute_coherence_score(real_top_words, dictionary=corpus_dictionary, texts=corpus.values, corpus=corpus_bow):.3f}\\n\")\n",
    "\n",
    "print(f\"Random words:\")\n",
    "for word_list in random_topics:\n",
    "    print(f\"    {word_list}\")\n",
    "print(f\"Coherence score for the text using random topics is: {utils.compute_coherence_score(random_topics, dictionary=corpus_dictionary, texts=corpus.values, corpus=corpus_bow):.3f}\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "If you run the previous cell multiple times you will get a different score for the random words (as different words are sampled with each run) but the score should always be around -20.\n",
    "\n",
    "As you would expect, there is a significant difference between actual top words of a topic and random words!"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 6. Topic Modeling with LDA\n",
    "### 6.1 Computing models with different number of topics\n",
    "\n",
    "Now that you have a way to compare the coherence of a collection of words, you can use this to measure the performance of LDA topic modeling algorithms by evaluating the coherence of the top words of each topic.\n",
    "You will now compute the coherence score for the LDA models ranging from 2 to 6 topics. But first you have to train the models:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Training LDA models from 2 to 6 topics\n",
      "Successfully trained LDA model on 2 topics\n",
      "Successfully trained LDA model on 3 topics\n",
      "Successfully trained LDA model on 4 topics\n",
      "Successfully trained LDA model on 5 topics\n",
      "Successfully trained LDA model on 6 topics\n"
     ]
    }
   ],
   "source": [
    "# Set the number of topics\n",
    "max_number_of_topics = 6\n",
    "n_topics_range = range(2, max_number_of_topics + 1)\n",
    "\n",
    "# Dictionary to save trained models\n",
    "LDA_models = {}\n",
    "print(f\"Training LDA models from 2 to {max_number_of_topics} topics\")\n",
    "\n",
    "# For number of topics ranging from 2 to max_number_of_topics\n",
    "for i in n_topics_range:\n",
    "    # Save model\n",
    "    LDA_models[i] = utils.train_lda_model(num_topics=i, corpus_bow=corpus_bow, corpus=corpus, corpus_dictionary=corpus_dictionary) # Training all models takes around 2 mins\n",
    "    print(f\"Successfully trained LDA model on {i} topics\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.2 Explore top words for different number of topics\n",
    "\n",
    "The coherence score metric is not perfect and within the subject of topic modeling there is no labeled ground truth, it all depends on what you are trying to accomplish. Have a look at the top words of each of the topics for models trained for different number of topics. While you can observe the coherence score as one metric of performance for different models, also look closely to get a sense of how human interpretation comes into this form of modeling, namely, despite whatever the coherence score says, can you get a sense of which models are producing more clearly relevant topics? Human interpretability is a key component of topic modeling in this way. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "c2b2c88719e44f429cb93bedc4e09276",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(Dropdown(description='# of topics', options=(2, 3, 4, 5, 6), value=2), Output()), _dom_c…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "utils.top_words_plot(LDA_models, max_number_of_topics)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 6.3 Compare coherence scores\n",
    "\n",
    "Now compute the coherence score for each model. Create a plot so it is easier to visualize:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 576x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# save coherence scores in a list\n",
    "coherence_scores_lda = [LDA_models[i].coherence_score for i in n_topics_range]\n",
    "\n",
    "# plot coherence score as a function of number of topics\n",
    "utils.plot_coherences_lda(n_topics_range, coherence_scores_lda)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 7. Select and evaluate the final LDA model\n",
    "According to the plot above, the optimal mumber of Topics is two (this may vary if you choose different words to exclude - see step 8). So you can select the model with two topics as your final model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 936x360 with 2 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "num_topics = 2\n",
    "\n",
    "# Select the model from the trained models\n",
    "lda_model_final = LDA_models[num_topics]\n",
    "\n",
    "# Plot the top ten words for the selected model again\n",
    "utils.plot_top_words_lda(num_topics, lda_model_final.model, num_words=20)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7.1 Analyze topic modeling of individual messages\n",
    "\n",
    "Now, you can use the model as something like a multiclass classification model. Here you will analyze one message at a time, and determine how well it belongs to each topic. "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "application/vnd.jupyter.widget-view+json": {
       "model_id": "d0a00b93ac3149698d4286c40f5ac456",
       "version_major": 2,
       "version_minor": 0
      },
      "text/plain": [
       "interactive(children=(Dropdown(description='Data to show', options=('All', 'medical_help', 'medical_products',…"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Here you can select a category of messages and look into what topic the single message would be classified\n",
    "utils.interact_with_filters(utils.plot_random_message_classification, haiti_df, corpus_dictionary=corpus_dictionary, lda_topics=lda_model_final)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### 7.2 Topic importance as a function of time\n",
    "Plot the topic importance over time in the next cell. For each of the topics the importance is calculated in each week after the disaster. You can see that the importance of some topics increases with time while it decreases for other topics. Try to see if this makes sense based on the top words of each topic."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1080x360 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "utils.plot_topic_importance(corpus_bow, haiti_df, lda_model_final, num_topics)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## 8. Removing additional words (optional)\n",
    "\n",
    "There are some words that are captured by the topic models as relevant because of their frequency, but are not really meaningful. In such cases, you can help the model by removing such words when you detect them. Examples of those words are 'like', 'would'. In the next cell you can include a list of words that the model must ignore. This is an important task where you can help the AI model. What other words do you identify? What do you think about 'please', 'need', and 'want'?"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Before removal:\n",
      "Dictionary contains a total of 1604 unique words\n",
      "\n",
      "\n",
      "After removal:\n",
      "Dictionary contains a total of 1602 unique words\n"
     ]
    }
   ],
   "source": [
    "# You can run this cell as many times as needed\n",
    "words_to_exclude = ['like', 'would']#, 'need', 'please'] # Add other words that you consider not important. \n",
    "print(f\"Before removal:\\nDictionary contains a total of {len(corpus_dictionary)} unique words\\n\")\n",
    "utils.exclude_words_from_dictionary(corpus_dictionary, words_to_exclude)\n",
    "\n",
    "# Create bag of words out of the full corpus\n",
    "corpus_bow = [corpus_dictionary.doc2bow(doc) for doc in corpus]\n",
    "# All words in the dictionary\n",
    "all_words = list(corpus_dictionary.values())\n",
    "print(f\"\\nAfter removal:\\nDictionary contains a total of {len(all_words)} unique words\")"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "When you remove the words you can return back to step 6 and try how the models perform with some of the words removed. If you want to reset your dictionary and BOW, you can return to step 3 and run the cells there to restore them."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Additional optional challenge / exploration\n",
    "\n",
    "**Look at topic modeling without removing any stopwords!**\n",
    "\n",
    "Run the cell below and then return to **step 3** above and run the cells in order to perform all the analysis in this lab without removing stopwords to see what you find. Here you'll read in the dataset again from scratch and then process the messages into tokens without removing stopwords. \n",
    "\n",
    "If you're a Python programmer, another thing you can try at this stage is looking at topic modeling for the other events captured in the dataset; Hurricane Sandy in the U.S. in 2012 or the Pakistan floods in 2010. You'll just need to change the filter below from `haiti_earthquake` to one of the other events, `usa_sandy` or `pakistan_floods`. But keep in mind the other events don't have dates associated with them so you'll need to comment out or modify any lines of code that are operating on the dates. \n",
    "\n",
    "Once you have loaded the data for a separate event, you'll have the option or training a new model for that event, or trying out your model trained on the Haiti data to see how it performs on another event. Try both if you're feeling ambitious and see what you find!"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "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>split</th>\n",
       "      <th>message</th>\n",
       "      <th>original</th>\n",
       "      <th>genre</th>\n",
       "      <th>related</th>\n",
       "      <th>PII</th>\n",
       "      <th>request</th>\n",
       "      <th>offer</th>\n",
       "      <th>aid_related</th>\n",
       "      <th>...</th>\n",
       "      <th>storm</th>\n",
       "      <th>fire</th>\n",
       "      <th>earthquake</th>\n",
       "      <th>cold</th>\n",
       "      <th>other_weather</th>\n",
       "      <th>direct_report</th>\n",
       "      <th>event</th>\n",
       "      <th>actionable_haiti</th>\n",
       "      <th>date_haiti</th>\n",
       "      <th>message_tokens</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>2</td>\n",
       "      <td>train</td>\n",
       "      <td>Weather update - a cold front from Cuba that c...</td>\n",
       "      <td>Un front froid se retrouve sur Cuba ce matin. ...</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-02-09</td>\n",
       "      <td>[weather, update, a, cold, front, from, cuba, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>7</td>\n",
       "      <td>train</td>\n",
       "      <td>Is the Hurricane over or is it not over</td>\n",
       "      <td>Cyclone nan fini osinon li pa fini</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>1</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[be, the, hurricane, over, or, be, it, not, over]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>12</td>\n",
       "      <td>train</td>\n",
       "      <td>says: west side of Haiti, rest of the country ...</td>\n",
       "      <td>facade ouest d Haiti et le reste du pays aujou...</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[say, west, side, of, haiti, rest, of, the, co...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>14</td>\n",
       "      <td>train</td>\n",
       "      <td>Information about the National Palace-</td>\n",
       "      <td>Informtion au nivaux palais nationl</td>\n",
       "      <td>direct</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[information, about, the, national, palace-]</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>15</td>\n",
       "      <td>train</td>\n",
       "      <td>Storm at sacred heart of jesus</td>\n",
       "      <td>Cyclone Coeur sacr de jesus</td>\n",
       "      <td>direct</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>...</td>\n",
       "      <td>1</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>0</td>\n",
       "      <td>haiti_earthquake</td>\n",
       "      <td>0</td>\n",
       "      <td>2010-01-17</td>\n",
       "      <td>[storm, at, sacred, heart, of, jesus]</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>5 rows × 46 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "   id  split                                            message  \\\n",
       "0   2  train  Weather update - a cold front from Cuba that c...   \n",
       "1   7  train            Is the Hurricane over or is it not over   \n",
       "2  12  train  says: west side of Haiti, rest of the country ...   \n",
       "3  14  train             Information about the National Palace-   \n",
       "4  15  train                     Storm at sacred heart of jesus   \n",
       "\n",
       "                                            original   genre  related  PII  \\\n",
       "0  Un front froid se retrouve sur Cuba ce matin. ...  direct        1    0   \n",
       "1                 Cyclone nan fini osinon li pa fini  direct        1    0   \n",
       "2  facade ouest d Haiti et le reste du pays aujou...  direct        1    0   \n",
       "3                Informtion au nivaux palais nationl  direct        0    0   \n",
       "4                        Cyclone Coeur sacr de jesus  direct        1    0   \n",
       "\n",
       "   request  offer  aid_related  ...  storm  fire  earthquake  cold  \\\n",
       "0        0      0            0  ...      0     0           0     0   \n",
       "1        0      0            1  ...      1     0           0     0   \n",
       "2        0      0            0  ...      0     0           0     0   \n",
       "3        0      0            0  ...      0     0           0     0   \n",
       "4        0      0            0  ...      1     0           0     0   \n",
       "\n",
       "   other_weather  direct_report             event  actionable_haiti  \\\n",
       "0              0              0  haiti_earthquake                 0   \n",
       "1              0              0  haiti_earthquake                 0   \n",
       "2              0              0  haiti_earthquake                 0   \n",
       "3              0              0  haiti_earthquake                 0   \n",
       "4              0              0  haiti_earthquake                 0   \n",
       "\n",
       "   date_haiti                                     message_tokens  \n",
       "0  2010-02-09  [weather, update, a, cold, front, from, cuba, ...  \n",
       "1  2010-01-17  [be, the, hurricane, over, or, be, it, not, over]  \n",
       "2  2010-01-17  [say, west, side, of, haiti, rest, of, the, co...  \n",
       "3  2010-01-17       [information, about, the, national, palace-]  \n",
       "4  2010-01-17              [storm, at, sacred, heart, of, jesus]  \n",
       "\n",
       "[5 rows x 46 columns]"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# OPTIONAL CHALLENGE / EXPLORATION: \n",
    "# Run the following lines of code to reload the data and tokenize without removing stopwords to see what happens\n",
    "# Load the datasets\n",
    "training_data = pd.read_csv(\"data/disaster_response_training.csv\", low_memory=False)\n",
    "validation_data = pd.read_csv(\"data/disaster_response_validation.csv\", low_memory=False)\n",
    "test_data = pd.read_csv(\"data/disaster_response_test.csv\", low_memory=False)\n",
    "# Merge the three datasets\n",
    "full_data = training_data.append(validation_data).append(test_data)\n",
    "# Fix column data type\n",
    "full_data['original'] = full_data['original'].astype(str)\n",
    "# Select only the Haiti data\n",
    "haiti_df = full_data[full_data.event == 'haiti_earthquake']\n",
    "# Fix column data type\n",
    "haiti_df['actionable_haiti'] = haiti_df.actionable_haiti.astype('int64')\n",
    "haiti_df['date_haiti'] = pd.to_datetime(haiti_df.date_haiti)\n",
    "# Get stopwords and punctuation from the utils file\n",
    "STOP_WORDS = [] #utils.STOP_WORDS\n",
    "punctuation = utils.punctuation\n",
    "# Instantiate the lemmatizer\n",
    "lemmatizer = WordNetLemmatizer()\n",
    "# Process all messages and save tokens to a new column\n",
    "haiti_df[\"message_tokens\"] = haiti_df.message.apply(\n",
    "    utils.process_text,\n",
    "    tokenizer=word_tokenize, pos_tagger=nltk.pos_tag, lemmatizer=lemmatizer, stopwords=STOP_WORDS, punctuation=punctuation\n",
    ")\n",
    "haiti_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}