{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Extracting TripAdvisor Reviews with Python"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Analyzing the \"voice-of-the-customer\" is the foundation of the market research industry. Typically, the industry has relied on two main methodologies and slight derivatives: surveys and focus groups. While it is unlikely that both of these methodologies will fall completely out of use, there are alternative techniques that utilize the power of modern analytics to extract valuable consumer information.\n",
    "\n",
    "The following is a brief exploration of using Python, Scrapy, the Natural Language Toolkit (NLTK), the **V**alence **A**ware **D**ictionary and s**E**ntiment **R**easoner (VADER) sentiment analyzer, and Scattertext, an NLP visualizer. The goal is to explore hotel and resort ratings for the Island of Hawai'i from TripAdvisor, understand consumer sentiment as it relates to different amenities, and visualize how reviews differ between positive and negative."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<style>.container { width:98% !important; }</style>"
      ],
      "text/plain": [
       "<IPython.core.display.HTML object>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "import pandas as pd\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import nltk\n",
    "from nltk.sentiment.vader import SentimentIntensityAnalyzer \n",
    "import scattertext as st\n",
    "import spacy\n",
    "import pickle\n",
    "from collections import Counter\n",
    "from IPython.display import IFrame\n",
    "from IPython.core.display import display, HTML\n",
    "%matplotlib inline\n",
    "display(HTML(\"<style>.container { width:98% !important; }</style>\"))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "First, we need to gather reviews. TripAdvisor (TA) is the number one travel site in the U.S. and boasts over 500 million hotel and resort reviews. For this demonstration, we are only interested in hotels/resorts on the Island of Hawai'i. Considering the island's heavy tourism industry, we have plenty of reviews to work with.\n",
    "\n",
    "The script below uses the Python package Scrapy which is designed specifically for web scraping tasks. To summarize, we provide an initial starting URL for the scraper -- in this case the first page of results when we search for *\"Island of Hawai'i\"* on TA-- and then use XPath references to extract review text from the underlying HTML code. This script first follows the link to a hotel's review page, extracts all reviews on the first page and subsequent pages, then goes back to the starting URL and follows the link to the next hotet, extracts all reviews...and so on until all reviews for all hotels have been scraped and added to an output CSV document.\n",
    "\n",
    "Referencing the function **\"parse_review\"** below, we've extracted the review headline, text, bubble rating, date, and hotel."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "import scrapy\n",
    "from tripadvisor.items import TripadvisorItem\n",
    "\n",
    "class TrSpider(scrapy.Spider):\n",
    "    name = 'trspider'\n",
    "    start_urls = [\n",
    "            'https://www.tripadvisor.com/Hotels-g29217-Island_of_Hawaii_Hawaii-Hotels.html'\n",
    "            ]\n",
    "\n",
    "    def parse(self, response):\n",
    "        for href in response.xpath('//div[@class=\"listing_title\"]/a/@href'):\n",
    "            url = response.urljoin(href.extract())\n",
    "            yield scrapy.Request(url, callback=self.parse_hotel)\n",
    "\n",
    "        next_page = response.xpath('//div[@class=\"unified pagination standard_pagination\"]/child::*[2][self::a]/@href')\n",
    "        if next_page:\n",
    "            url = response.urljoin(next_page[0].extract())\n",
    "            yield scrapy.Request(url, self.parse)\n",
    "\n",
    "    def parse_hotel(self, response):\n",
    "        for href in response.xpath('//div[starts-with(@class,\"quote\")]/a/@href'):\n",
    "            url = response.urljoin(href.extract())\n",
    "            yield scrapy.Request(url, callback=self.parse_review)\n",
    "\n",
    "\n",
    "        next_page = response.xpath('//link[@rel=\"next\"]/@href')\n",
    "        if next_page:\n",
    "            url = response.urljoin(next_page[0].extract())\n",
    "            yield scrapy.Request(url, self.parse_hotel)\n",
    "\n",
    "    def parse_review(self, response):\n",
    "        item = TripadvisorItem()\n",
    "        item['headline'] = response.xpath('translate(//div[@class=\"quote\"]/text(),\"!\",\" \")').extract()[0][1:-1]\n",
    "        item['review'] = response.xpath('translate(//div[@class=\"entry\"]/p,\"\\n\",\" \")').extract()[0]\n",
    "        item['bubbles'] = response.xpath('//span[contains(@class,\"ui_bubble_rating\")]/@alt').extract()[0]\n",
    "        item['date'] = response.xpath('normalize-space(//span[contains(@class,\"ratingDate\")]/@content)').extract()[0]\n",
    "        item['hotel'] = response.xpath('normalize-space(//span[@class=\"altHeadInline\"]/a/text())').extract()[0]\n",
    "        return item"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The raw output generates a file with **~41,500** records spanning the dates of **2003-01-20** to **2017-07-20** -- over 14 years of reviews! Before we start exploring though we'll need to run some data cleaning steps. The function below loads the CSV into Python, removes reviews with missing values, and drops hotels with fewer than 100 reviews."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def clean_scrapy(raw_csv):\n",
    "    df = pd.read_csv(raw_csv, dtype={'bubbles':'category','hotel':'category'}, parse_dates=[1])\n",
    "    df.dropna(inplace=True)\n",
    "    df.bubbles.cat.rename_categories([1,2,3,4,5], inplace=True)\n",
    "    df = df.groupby('hotel').filter(lambda x: len(x) > 100)\n",
    "    df['hotel'] = df['hotel'].cat.remove_unused_categories()\n",
    "    return df\n",
    "\n",
    "df = clean_scrapy('bihotels.csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "This operation reduces our data set to **~40,800** reviews.\n",
    "\n",
    "Since this is mostly an exploratory exercise, we'd like a means of slicing and dicing our data to look at it in different ways. The function below allows us to filter by review dates, bubble rating counts, and hotel. It also lets us return review headlines rather than the review itself should we want to explore those too. For now, we'll just look at all the reviews."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {
    "collapsed": true,
    "scrolled": true
   },
   "outputs": [],
   "source": [
    "mindate = df['date'].min()\n",
    "maxdate = df['date'].max()\n",
    "def data_filter(df, min_date=mindate, max_date=maxdate, low_bub=1, high_bub=5, hotel=None, headline=False):\n",
    "    text = df\n",
    "    text = text.query('date >= @min_date & date <= @max_date')\n",
    "    text = text.query('bubbles >= @low_bub & bubbles <= @high_bub')\n",
    "    if hotel:\n",
    "        text = text.query('hotel == @hotel')\n",
    "    return text[['bubbles','headline']] if headline else text[['bubbles','review']]\n",
    "\n",
    "text = data_filter(df)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Natural Language Processing (NLP) is an entirely vast subfield of machine learning (ML) if not a field in its own right. The variety of analyses and applications is (in my opinion) much greater than many other ML pursuits. Additionally, when considering nuance, context, and semantic idiosyncrasies of written and spoke language NLP is one of the most complex areas in ML.\n",
    "\n",
    "Three of the most common NLP preprocessing steps are case folding, tokenization, and part-of-speech tagging:\n",
    "\n",
    "* **Case Folding** is the process of transforming text into lowercase. This is mainly done so that words with difference cases are not treated as separate words by certain NLP algorithms (e.g. \"Resort\" and \"resort\" could be seen as two different words).\n",
    "\n",
    "\n",
    "* **Tokenization** takes each word in a collection of text (referred to as a \"corpus\" in NLP terms) and assigns it a unique numeric value. Words themselves are not numeric or quantitative in nature but many of the algorithms used in NLP require them to be treated that way. This is another reason for Case Folding so we don't have words with different assignments (e.g. \"What a great hotel!\" tokenized would be \"What\",\"a\",\"great\",\"hotel\",\"!\". Punctuation is optional depending on application).\n",
    "\n",
    "\n",
    "* **Part-of-Speech Tagging** assigns words to word types like \"nouns,\" \"verbs,\" adjectives,\" etc. This makes it easier to extract informative components of sentences like noun phrases (e.g. \"wonderful spa\", \"beautiful pool\", \"horrible restaurant\" are all examples of simple noun phrases -- an adjective followed by a noun)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def np_preprocess(text):\n",
    "    corpus = ' '.join(text.iloc[:,1])\n",
    "    sentences = nltk.sent_tokenize(corpus.lower())\n",
    "    tokens = [nltk.word_tokenize(token) for token in sentences] \n",
    "    tags = [nltk.pos_tag(tag) for tag in tokens]\n",
    "    return tags\n",
    "\n",
    "tags = np_preprocess(text)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Identifying noun phrases will be our first task. The rationale is that when a consumer writes a review on TA they are given on option to also rate things like \"Value\" and \"Sleep Quality\" but they don't always do. However, that doesn't mean that their review itself doesn't explicitly mention these types of things or specific amenities. We can potentially get a clearer picture using NLP techniques.\n",
    "\n",
    "The script below takes our review corpus (each word having been tagged with a part-of-speech from our previous step) and extracts the nouns. The idea is that commonly occurring nouns are a good indicator of hotel/resort features and amenities that reviewers feel strongly about. It gives us a good starting place for sentiment analysis"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {
    "scrolled": false
   },
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4sAAAJCCAYAAABpg1KrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2689065b748>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "nouns = []\n",
    "for tag in tags:\n",
    "    for np in tag:\n",
    "        if np[1] in ['NN','NNP','NNS','NNPS']: #This covers all noun types, normal, proper, and plural\n",
    "            nouns.append(np)\n",
    "\n",
    "plt.figure(figsize=(15,8))\n",
    "nltk.FreqDist(nouns).plot(50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The chart above shows the 50 most commonly occurring nouns in our review corpus. This part requires some subjectivity to determine which nouns are worth exploring more. The first five (\"i,\" \"hotel,\" \"room,\" \"beach,\" \"resort\") are general and might not be of much help but \"pool,\" \"staff, \"food,\" and \"service\" might warrant further investigation. \n",
    "\n",
    "In a real-word setting, we'd probably want to work with managers and industry experts to tease out the best nouns to use. We may also want to consider the time of reference since some reviews may have terse comments regarding something the resort addressed recently (i.e. remodeled pool or improved restaurant menu).\n",
    "\n",
    "Let's take a deeper look at \"pool,\" \"staff,\" \"food,\" \"service,\" and we'll add in \"location\" as well since that is one feature that TA also tracks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## VADER Sentiment Analyzer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "Understanding the positive and negative emotions consumers have towards a product or service is of significant interest to businesses. While market researchers typically approach these types of problems with survey and focus group methodology, NLP provides some alternative methods that 1) could provide less biased insights and, 2) can cost far less than the price of recruiting survey respondents and focus group participants.\n",
    "\n",
    "Sentiment Analysis is the NLP approach to uncovering these emotions. While being one of the more challenging NLP tasks, there are a variety of methodologies. We chose to use VADER for this demonstration. You can read more about it [here](http://comp.social.gatech.edu/papers/icwsm14.vader.hutto.pdf). Many market researchers readily say that \"machines\" will never replace humans in understanding other humans, yet VADER itself was actually trained *using humans to score sentiment polarity!*\n",
    "\n",
    "Some interesting features of the VADER Sentiment Analyzer:\n",
    "\n",
    "* Considers punctuation as an intensity modifier, so a sentence ending in **\"!\"** isn't as intense as one ending with **\"!!\"**\n",
    "* Considers ALL CAPS as an intensity modifier, so **\"great\"** isn't as intense as **\"GREAT.\"**\n",
    "* Considers adjectives and other word modifiers, so **\"It's good\"** isn't as intense as **\"It's really good.\"**\n",
    "* Considers conjunctive sentences using \"but\" as two different sentiments with more weight placed on sentiment after \"but.\"\n",
    "* Considers tri-grams (i.e. groups of three words together in a sentence) for sentiment \"flips\" as in **\"This resort really isn't all that great.\"**"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "sia = nltk.sentiment.SentimentIntensityAnalyzer()\n",
    "\n",
    "def feature_sentiments(feature):\n",
    "    '''\n",
    "    Some tagged words include a '.' if reviewer didn't add spacing after \n",
    "    end-of-sentence period making parser include last word of sentence with \n",
    "    first word of next(e.g. 'hotel.but'). Causes division by zero since no \n",
    "    sentence contains these terms by themselves\n",
    "    '''    \n",
    "    feature = [sentence + '.' for sentence in text.iloc[:,1] if feature in sentence]\n",
    "    sentiment = []\n",
    "    for sentence in feature:       \n",
    "        sentiment.append(sia.polarity_scores(sentence))\n",
    "\n",
    "    if len(sentiment) != 0:            \n",
    "        results = sum([item['compound'] for item in sentiment])/len(sentiment)  \n",
    "    else:\n",
    "        results = 0\n",
    "    return results"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The VADER algorithm takes a sentence as input and compares it to a lexicon created during the initial algorithm training by humans. Each sentence is given positive, neutral, negative, and compound score based on reference to this lexicon and with the 5 bullets above. We will use just the compound score since it is an amalgamation of sentiment scores for each word in a sentence normalized between -1 and 1 with \"-1\" being \"completely negative\" and \"1\" being \"completely positive.\" This is convenient since we can then set our own thresholds for what we believe constitutes positive/neutral/negative.\n",
    "\n",
    "The above script takes a key word, extracts all sentences from the review corpus that contain it, scores them on sentiment, then averages the compound score across all sentences mentioning the key word(s). We decided to look at the words \"pool,\" \"staff,\" \"food,\" \"service,\" and \"location.\""
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>pool</th>\n",
       "      <td>0.879040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>staff</th>\n",
       "      <td>0.877913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>food</th>\n",
       "      <td>0.875209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>service</th>\n",
       "      <td>0.843583</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>location</th>\n",
       "      <td>0.873114</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "                 0\n",
       "pool      0.879040\n",
       "staff     0.877913\n",
       "food      0.875209\n",
       "service   0.843583\n",
       "location  0.873114"
      ]
     },
     "execution_count": 9,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "features = {'pool': '','staff': '','food': '','service': '','location': ''}\n",
    "for feature in features.keys():\n",
    "    features[feature] = feature_sentiments(feature)    \n",
    "pd.DataFrame.from_dict(features, orient='index')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "So what we can see here is that most mentions of these terms are positive (remember we didn't filter the data by bubble rating so both negative and positive reviews are presented). But let's dig a little deeper: *why* do reviewers rate these positive? The NLTK also offers a few other tools for NLP that might be helpful. The \"concordance\" method allows us to choose a key word and see its context within sentences. This can add additional \"flavor\" to the reason reviewers provided the rating they did."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "def concord (text, feature):\n",
    "    corpus = ' '.join(text.iloc[:,1])\n",
    "    words = nltk.word_tokenize(corpus.lower())\n",
    "    concordance = nltk.Text(words)\n",
    "    return concordance.concordance(feature) "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Displaying 25 of 24723 matches:\n",
      "ntastic . we enjoyed our time at the pool . it was a quick stay and would defi\n",
      "ying to everything on the cheap from pool food service hours to the fairmont b\n",
      "s that has truly everything ! lovely pool , great waterview from many rooms , \n",
      " can snorkel with , a huge freestyle pool , staff which is super friendly and \n",
      "nt you can rent beach / beach-pool / pool cabanas which come with some snacks \n",
      " , grab something in the deli at the pool or get a ride over to the farmlands \n",
      "g is beautiful . you may swim in the pool or in the lagoon . there is an equip\n",
      " snorkeling , paddle boarding , tide pool exploring and watching the turtles !\n",
      "softer side and felt a little worn . pool was out of service for half our stay\n",
      " rooms havebalconies overlooking the pool and harbor . the japanese gardens ar\n",
      "ach , a nesting turtles area . great pool , with lots of chaise lounges . fant\n",
      " showers ... .one with a tub . large pool with lots of kids running around ...\n",
      "ounger daughter especially loved the pool and water slide . we all enjoyed the\n",
      "d area of lounge chairs right by the pool . every staff member we encountered \n",
      "ng housekeeping for service . 4. the pool was under renovation . fortunatly , \n",
      "d did n't have time to check out the pool or the tide pool area , but it looke\n",
      "me to check out the pool or the tide pool area , but it looked gorgeous.howeve\n",
      "t the front desk . we discovered the pool was closed ; very bad news for my si\n",
      " to using it . we were told that the pool was closed for regular maintenance b\n",
      "oom was clean and had a balcony with pool view . the staff was friendly and he\n",
      "ncredible sunset dinner . all of the pool staff and waitresses who served us a\n",
      " waitresses who served us around the pool were very friendly . all of the vale\n",
      "vice and taste ) at the brown at the pool bar and breakfast ran out of mint on\n",
      "cancelled . no drinks service by the pool on 7/2 drink service closed at 4 on \n",
      " 'm returning from a long day at the pool or beach , my room still is n't clea\n"
     ]
    }
   ],
   "source": [
    "concord(text,'pool')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "For example, if we explore the word \"pool\" we can see more context around reviewers' experience with the resort's pool: \n",
    "\n",
    "* **\"lovely *pool*, great waterview from many rooms\"**\n",
    "* **\"*pool* was out of service for half our stay\"**\n",
    "* **\"loved the *pool* and water slide\"**\n",
    "\n",
    "Remember, we are using the full, unfiltered data set so these comments would probably be more helpful if we narrowed our data to a specific resort and/or time span."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## ...but what about just the reviews for, say, the Hilton Waikoloa?"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Remember, we are looking at the full set of reviews we scraped from TA. In a real world environment, an operations manager would probably be mostly concerned with their own resort's reviews. Being one of the larger resorts on Hawai'i Island, the Hilton Waikoloa also has several thousand reviews we can examine (~7,000 over our 14 year time period)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 12,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA4sAAAJCCAYAAABpg1KrAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\nAAAASUVORK5CYII=\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x26883671da0>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "hilton_reviews = data_filter(df, hotel='Hilton Waikoloa Village')\n",
    "hilton_tags = np_preprocess(hilton_reviews)\n",
    "\n",
    "hilton_nouns = []\n",
    "for tag in hilton_tags:\n",
    "    for np in tag:\n",
    "        if np[1] in ['NN','NNP','NNS','NNPS']: #This covers all noun types, normal, proper, and plural\n",
    "            hilton_nouns.append(np)\n",
    "\n",
    "plt.figure(figsize=(15,8))\n",
    "nltk.FreqDist(nouns).plot(50)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hilton also receives mostly positive reviews:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.axes._subplots.AxesSubplot at 0x2689e60a390>"
      ]
     },
     "execution_count": 13,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "iVBORw0KGgoAAAANSUhEUgAAA3cAAAHeCAYAAAA4v/IZAAAABHNCSVQICAgIfAhkiAAAAAlwSFlz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\n",
      "text/plain": [
       "<matplotlib.figure.Figure at 0x2689065b128>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "hilton_reviews['bubbles'].value_counts().plot(kind='bar', figsize=(15,8), title='Hilton Waikoloa Bubble Rating Counts')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As we did earlier, let's look at nouns that are mentioned most often and dive deeper into sentiments surrounding each. For this example we'll use \"lagoon\" (the hotel sits on a lagoon guests can use for water activities), \"food,\" \"pool,\" \"staff,\" and \"tram\" (because of the hotel's size, a passenger tram operates all day to shuttle guests from one end to the other). Both \"lagoon\" and \"tram\" may be interesting to look at since they are unique to the Hilton."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>0</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>lagoon</th>\n",
       "      <td>0.897359</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>food</th>\n",
       "      <td>0.875209</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>pool</th>\n",
       "      <td>0.879040</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>staff</th>\n",
       "      <td>0.877913</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>tram</th>\n",
       "      <td>0.834116</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "               0\n",
       "lagoon  0.897359\n",
       "food    0.875209\n",
       "pool    0.879040\n",
       "staff   0.877913\n",
       "tram    0.834116"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hilton_features = {'lagoon': '','food': '','pool': '','staff': '','tram': ''}\n",
    "for feature in hilton_features.keys():\n",
    "    hilton_features[feature] = feature_sentiments(feature)    \n",
    "pd.DataFrame.from_dict(hilton_features, orient='index')"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Like we saw before when looking at all reviews, nouns with the most mentions tend to be in a positive context. This time, rather than cherry-picking let's review *all* the extracted nouns (that were mentioned more than, say, 100 times) to find which ones are also associated mostly with negative sentiments. This gives us **~300** nouns to review."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "'''\n",
    "Commented-out script was ran initially to capture VADER sentiment scores.\n",
    "Scores were stored in a pickle (*.pkl) file to avoid rerunning the script which\n",
    "can take several hours to complete.\n",
    "'''\n",
    "\n",
    "#hilton_nouns = [e for e in Counter(word[0] for word in hilton_nouns).most_common() if e[1] > 100]\n",
    "#all_hilton_features = dict([(noun, '') for noun,value in hilton_nouns])\n",
    "#for feature in all_hilton_features.keys():\n",
    "#    all_hilton_features[feature] = feature_sentiments(feature)\n",
    "    \n",
    "#with open('hilton_sentiments.pkl', \"wb\") as file:\n",
    "#    pickle.dump(all_hilton_features, file , protocol=0)\n",
    "    \n",
    "with open('hilton_sentiments.pkl', \"rb\") as file:\n",
    "    all_hilton_features = pickle.load(file)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style>\n",
       "    .dataframe thead tr:only-child th {\n",
       "        text-align: right;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: left;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>Noun</th>\n",
       "      <th>Sentiment</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>kpc</td>\n",
       "      <td>0.984700</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>beauty</td>\n",
       "      <td>0.961632</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>sunsets</td>\n",
       "      <td>0.955682</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>kamuela</td>\n",
       "      <td>0.955133</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>paradise</td>\n",
       "      <td>0.952219</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>turtle</td>\n",
       "      <td>0.950189</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>snacks</td>\n",
       "      <td>0.949627</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>turtles</td>\n",
       "      <td>0.949606</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>waterfall</td>\n",
       "      <td>0.948536</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>relaxing</td>\n",
       "      <td>0.947408</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>10</th>\n",
       "      <td>waterfalls</td>\n",
       "      <td>0.947134</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>11</th>\n",
       "      <td>sunset</td>\n",
       "      <td>0.944992</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>12</th>\n",
       "      <td>kayaks</td>\n",
       "      <td>0.944483</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>13</th>\n",
       "      <td>sandwiches</td>\n",
       "      <td>0.944418</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>14</th>\n",
       "      <td>disneyland</td>\n",
       "      <td>0.943989</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>15</th>\n",
       "      <td>snorkeling</td>\n",
       "      <td>0.943084</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>16</th>\n",
       "      <td>artwork</td>\n",
       "      <td>0.943069</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>17</th>\n",
       "      <td>sushi</td>\n",
       "      <td>0.942973</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>18</th>\n",
       "      <td>helicopter</td>\n",
       "      <td>0.942122</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>19</th>\n",
       "      <td>plenty</td>\n",
       "      <td>0.941598</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>20</th>\n",
       "      <td>fish</td>\n",
       "      <td>0.941107</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>21</th>\n",
       "      <td>sports</td>\n",
       "      <td>0.941103</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>22</th>\n",
       "      <td>snorkel</td>\n",
       "      <td>0.941030</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>23</th>\n",
       "      <td>toys</td>\n",
       "      <td>0.940731</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>24</th>\n",
       "      <td>variety</td>\n",
       "      <td>0.940406</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>25</th>\n",
       "      <td>statues</td>\n",
       "      <td>0.939203</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>26</th>\n",
       "      <td>grocery</td>\n",
       "      <td>0.938562</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>27</th>\n",
       "      <td>whales</td>\n",
       "      <td>0.937186</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>28</th>\n",
       "      <td>highlight</td>\n",
       "      <td>0.936739</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>29</th>\n",
       "      <td>advantage</td>\n",
       "      <td>0.936449</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>...</th>\n",
       "      <td>...</td>\n",
       "      <td>...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>516</th>\n",
       "      <td>bill</td>\n",
       "      <td>0.774680</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>517</th>\n",
       "      <td>bags</td>\n",
       "      <td>0.773301</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>518</th>\n",
       "      <td>nothing</td>\n",
       "      <td>0.772597</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>519</th>\n",
       "      <td>desk</td>\n",
       "      <td>0.772015</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>520</th>\n",
       "      <td>marriott</td>\n",
       "      <td>0.765239</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>521</th>\n",
       "      <td>call</td>\n",
       "      <td>0.762300</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>522</th>\n",
       "      <td>lady</td>\n",
       "      <td>0.762280</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>523</th>\n",
       "      <td>system</td>\n",
       "      <td>0.758163</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>524</th>\n",
       "      <td>someone</td>\n",
       "      <td>0.755727</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>525</th>\n",
       "      <td>elevators</td>\n",
       "      <td>0.755673</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>526</th>\n",
       "      <td>problem</td>\n",
       "      <td>0.752245</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>527</th>\n",
       "      <td>dollars</td>\n",
       "      <td>0.750019</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>528</th>\n",
       "      <td>charges</td>\n",
       "      <td>0.745355</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>529</th>\n",
       "      <td>elevator</td>\n",
       "      <td>0.744285</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>530</th>\n",
       "      <td>credit</td>\n",
       "      <td>0.743032</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>531</th>\n",
       "      <td>charge</td>\n",
       "      <td>0.742932</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>532</th>\n",
       "      <td>months</td>\n",
       "      <td>0.737567</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>533</th>\n",
       "      <td>phone</td>\n",
       "      <td>0.728441</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>534</th>\n",
       "      <td>customer</td>\n",
       "      <td>0.723255</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>535</th>\n",
       "      <td>maintenance</td>\n",
       "      <td>0.708988</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>536</th>\n",
       "      <td>security</td>\n",
       "      <td>0.699416</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>537</th>\n",
       "      <td>problems</td>\n",
       "      <td>0.694774</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>538</th>\n",
       "      <td>management</td>\n",
       "      <td>0.685590</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>539</th>\n",
       "      <td>manager</td>\n",
       "      <td>0.661050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>540</th>\n",
       "      <td>hhonors</td>\n",
       "      <td>0.661050</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>541</th>\n",
       "      <td>carpet</td>\n",
       "      <td>0.659260</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>542</th>\n",
       "      <td>disappointment</td>\n",
       "      <td>0.637888</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>543</th>\n",
       "      <td>smell</td>\n",
       "      <td>0.636388</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>544</th>\n",
       "      <td>oahu</td>\n",
       "      <td>0.534360</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>545</th>\n",
       "      <td>kirin</td>\n",
       "      <td>-0.571900</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "<p>546 rows × 2 columns</p>\n",
       "</div>"
      ],
      "text/plain": [
       "               Noun  Sentiment\n",
       "0               kpc   0.984700\n",
       "1            beauty   0.961632\n",
       "2           sunsets   0.955682\n",
       "3           kamuela   0.955133\n",
       "4          paradise   0.952219\n",
       "5            turtle   0.950189\n",
       "6            snacks   0.949627\n",
       "7           turtles   0.949606\n",
       "8         waterfall   0.948536\n",
       "9          relaxing   0.947408\n",
       "10       waterfalls   0.947134\n",
       "11           sunset   0.944992\n",
       "12           kayaks   0.944483\n",
       "13       sandwiches   0.944418\n",
       "14       disneyland   0.943989\n",
       "15       snorkeling   0.943084\n",
       "16          artwork   0.943069\n",
       "17            sushi   0.942973\n",
       "18       helicopter   0.942122\n",
       "19           plenty   0.941598\n",
       "20             fish   0.941107\n",
       "21           sports   0.941103\n",
       "22          snorkel   0.941030\n",
       "23             toys   0.940731\n",
       "24          variety   0.940406\n",
       "25          statues   0.939203\n",
       "26          grocery   0.938562\n",
       "27           whales   0.937186\n",
       "28        highlight   0.936739\n",
       "29        advantage   0.936449\n",
       "..              ...        ...\n",
       "516            bill   0.774680\n",
       "517            bags   0.773301\n",
       "518         nothing   0.772597\n",
       "519            desk   0.772015\n",
       "520        marriott   0.765239\n",
       "521            call   0.762300\n",
       "522            lady   0.762280\n",
       "523          system   0.758163\n",
       "524         someone   0.755727\n",
       "525       elevators   0.755673\n",
       "526         problem   0.752245\n",
       "527         dollars   0.750019\n",
       "528         charges   0.745355\n",
       "529        elevator   0.744285\n",
       "530          credit   0.743032\n",
       "531          charge   0.742932\n",
       "532          months   0.737567\n",
       "533           phone   0.728441\n",
       "534        customer   0.723255\n",
       "535     maintenance   0.708988\n",
       "536        security   0.699416\n",
       "537        problems   0.694774\n",
       "538      management   0.685590\n",
       "539         manager   0.661050\n",
       "540         hhonors   0.661050\n",
       "541          carpet   0.659260\n",
       "542  disappointment   0.637888\n",
       "543           smell   0.636388\n",
       "544            oahu   0.534360\n",
       "545           kirin  -0.571900\n",
       "\n",
       "[546 rows x 2 columns]"
      ]
     },
     "execution_count": 16,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "pd.DataFrame(sorted(all_hilton_features.items(), key=lambda x: x[1], reverse=True), columns=('Noun','Sentiment'))"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Scattertext Visualizer"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "As with most pursuits in data exploration, visualization is a helpful tool. NLP is no different. In our exploration of TA reviews we might be interested in what words or phrases best differentiate between positive and negative reviews. For this task we'll look at the bubble ratings reviewers left and define positive reviews as 5 out of 5 bubbles and negative reviews as 1 out of 5 bubbles. We chose this dichotomy since experience in market research shows that people tend to give ratings closer to the middle of a scale unless they feel very strongly one way or the other. Again, the end user may want to use different thresholds.\n",
    "\n",
    "Scattertext is an open-source library for Python for visualizing how word usage differs between two groups. The process uses a modified version of F-Score calculation (i.e. the harmonic mean of precision and frequency) called \"Scaled F-Score\" to partition words by groups (you can read more about it and the Scattertext package [here](https://github.com/JasonKessler/scattertext)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "metadata": {
    "collapsed": true
   },
   "outputs": [],
   "source": [
    "neg = data_filter(df, high_bub=1)\n",
    "pos = data_filter(df, low_bub=5)\n",
    "\n",
    "scatter_reviews = pd.concat([pos,neg])\n",
    "scatter_reviews['bubbles'] = scatter_reviews.bubbles.astype(str)\n",
    "\n",
    "spacy_corpus = spacy.en.English()\n",
    "scatter_corpus = st.CorpusFromPandas(scatter_reviews, \n",
    "                                     category_col='bubbles', \n",
    "                                     text_col='review', \n",
    "                                     nlp=spacy_corpus).build()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "The code above divides our reviews into groups for 1 bubble and 5 bubbles. Each set of reviews is parsed and compared by raw frequency within its category and its frequency across both categories. This effectively scores each word on how representative it is of a 1 or 5 bubble review (again see the Scaled F-Score referenced in the link above).\n",
    "\n",
    "In the visualization below, each axis represents a review and each point a word. Words in the upper left represent words used more frequently by 5 bubble reviews while bottom right word are more frequently used by 1 bubble reviews. Words that appear in the middle could appear in either so they aren't as differentiating between groups. On the right, we are provided a list of the top words characteristic to each group and characteristic to the entire corpus of text (note: in the given code, words with fewer than 30 mentions are excluded).\n",
    "\n",
    "A quick review shows that the use of adjectives is really what sets these reviews apart. While not particularly helpful, the words in the \"Characteristic\" column can give a little more information on what exactly is \"superb\" or \"disgusting.\" This improves interpretability marginally, but clicking on a word within the visualization will create a list or reviews for each group containing that word. This makes it much easier to do a deeper exploration. Finally, a user can enter their own word to search for in the corpus to see how it is used by each group.  "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "        <iframe\n",
       "            width=\"1300\"\n",
       "            height=\"700\"\n",
       "            src=\"posneg_scatter.html\"\n",
       "            frameborder=\"0\"\n",
       "            allowfullscreen\n",
       "        ></iframe>\n",
       "        "
      ],
      "text/plain": [
       "<IPython.lib.display.IFrame at 0x26885178ef0>"
      ]
     },
     "execution_count": 18,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "html = st.produce_scattertext_explorer(scatter_corpus,\n",
    "                                       category='5',\n",
    "                                       category_name='5 of 5 bubbles',\n",
    "                                       not_category_name='1 of 5 bubbles',\n",
    "                                       minimum_term_frequency = 30,\n",
    "                                       save_svg_button=True)\n",
    "open(\"posneg_scatter.html\", 'wb').write(html.encode('utf-8'))\n",
    "IFrame('posneg_scatter.html', width=1300, height=700)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "We'll do the same for just Hilton Waikoloa:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 21,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/html": [
       "\n",
       "        <iframe\n",
       "            width=\"1300\"\n",
       "            height=\"700\"\n",
       "            src=\"hilton_posneg_scatter.html\"\n",
       "            frameborder=\"0\"\n",
       "            allowfullscreen\n",
       "        ></iframe>\n",
       "        "
      ],
      "text/plain": [
       "<IPython.lib.display.IFrame at 0x2699c46f780>"
      ]
     },
     "execution_count": 21,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "hilton_neg = data_filter(df, high_bub=1, hotel='Hilton Waikoloa Village')\n",
    "hilton_pos = data_filter(df, low_bub=5, hotel='Hilton Waikoloa Village')\n",
    "\n",
    "hilton_scatter_reviews = pd.concat([hilton_pos,hilton_neg])\n",
    "hilton_scatter_reviews['bubbles'] = hilton_scatter_reviews.bubbles.astype(str)\n",
    "\n",
    "hilton_spacy_corpus = spacy.en.English()\n",
    "hilton_scatter_corpus = st.CorpusFromPandas(hilton_scatter_reviews, \n",
    "                                            category_col='bubbles', \n",
    "                                            text_col='review', \n",
    "                                            nlp=hilton_spacy_corpus).build()\n",
    "\n",
    "hilton_html = st.produce_scattertext_explorer(hilton_scatter_corpus,\n",
    "                                              category='5',\n",
    "                                              category_name='5 of 5 bubbles',\n",
    "                                              not_category_name='1 of 5 bubbles',\n",
    "                                              minimum_term_frequency = 30,\n",
    "                                              save_svg_button=True)\n",
    "open(\"hilton_posneg_scatter.html\", 'wb').write(hilton_html.encode('utf-8'))\n",
    "IFrame('hilton_posneg_scatter.html', width=1300, height=700)"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {
    "collapsed": true
   },
   "source": [
    "## Final Remarks"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "In the body of NLP analytics, there are numerous paths one can take in their analysis. Additional applications could include:\n",
    "\n",
    "* Being the most popular travel review site online, TA is generating reviews 24/7 from around the world. Creating a data pipeline that regularly collects reviews for processing can be a part of an ongoing customer service effort for a single hotel, a hospitality holding company managing numerous properties, or even a tourism board tasked with keeping visitors' experiences memorable.\n",
    "\n",
    "* The Scattertext visualization also allows for the hotel name to be attached to each review. While the above code does not include this feature, in a real-world use case it would be imperative to know if a good/bad review was given for your hotel or a competitor's.\n",
    "\n",
    "* Businesses spend millions annually on market research projects that often provide little value for the cost. Focus groups are a mainstay of the corporate research body of methodologies and attempt to extract similar information. By using online reviews it may be possible to achieve the same level of insight for little to no cost, from thousands of reviewers, and on an ongoing basis."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Citations"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Hutto, C.J. & Gilbert, E.E. (2014). VADER: A Parsimonious Rule-based Model for Sentiment Analysis of Social Media Text. Eighth International Conference on Weblogs and Social Media (ICWSM-14). Ann Arbor, MI, June 2014.\n",
    "\n",
    "Jason S. Kessler. Scattertext: a Browser-Based Tool for Visualizing how Corpora Differ. Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (ACL): System Demonstrations. 2017."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python [conda root]",
   "language": "python",
   "name": "conda-root-py"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.6.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}