Mining-Yelp-Reviews-with-Topic-Modeling-Technique.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "93f0ed66",
   "metadata": {},
   "source": [
    "### Mining Consumer Reviews for Insights using Topic Modelling for Short Text\n",
    "\n",
    "\n",
    "\n",
    "- Mining Yelp Reviews for Insights. Yelp dataset is availaible from https://www.yelp.com/dataset in json. Be warned that it's about 4gb file\n",
    "- We will break the reviews down into sentences and cluster them using the gsdmm package. The resulting clusters should be about similar aspects and experience, and while many reviews are about restaurants, there are also other reviews, such as those concerning nail salon ratings.\n",
    "\n",
    "#### Packages Required\n",
    "\n",
    "Install GSDM git clone https://github.com/rwalk/gsdmm.git it's a handy package for Topic modelling for short text"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "id": "dfa79514",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "id": "c0a85c64",
   "metadata": {},
   "outputs": [],
   "source": [
    "#Import mgp from GSDM \n",
    "\n",
    "from mgp import MovieGroupProcess"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "id": "bb65cfe3",
   "metadata": {},
   "outputs": [],
   "source": [
    "import string\n",
    "import pickle"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "id": "7b448989",
   "metadata": {},
   "outputs": [],
   "source": [
    "from phrases import get_yelp_reviews"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "id": "e6f2182c",
   "metadata": {},
   "outputs": [],
   "source": [
    "from preprocess_bbc_dataset import get_stopwords"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "id": "d60ba8e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "import nltk"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "id": "8e030a97",
   "metadata": {},
   "outputs": [],
   "source": [
    "tokenizer = nltk.data.load(\"tokenizers/punkt/english.pickle\")\n",
    "yelp_reviews_file = \"yelp_academic_dataset_review.json\"\n",
    "stopwords_file_path = \"reviews_stopwords.csv\"\n",
    "stopwords = get_stopwords(stopwords_file_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 8,
   "id": "d5e00b77",
   "metadata": {},
   "outputs": [],
   "source": [
    "def preprocess(text):\n",
    "    sentences = tokenizer.tokenize(text)\n",
    "    sentences = [nltk.tokenize.word_tokenize(sentence) for sentence in sentences]\n",
    "    sentences = [list(set(word_list)) for word_list in sentences]\n",
    "    sentences = [[word for word in word_list if word not in stopwords and word not in string.punctuation] for word_list in sentences]\n",
    "    return sentences"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "205fae0e",
   "metadata": {},
   "source": [
    "we define the preprocessing function. This function first splits the text into sentences, tokenizes the sentences into words, and removes duplicates from the word lists. The duplicate removal is necessary for the GSDMM model, as it requires a list of unique tokens that occur in the text. The preprocessing function then removes stopwords and punctuation from the word lists.\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 9,
   "id": "1e0d8c6e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def top_words_by_cluster(mgp, top_clusters, num_words):\n",
    "    for cluster in top_clusters:\n",
    "        sort_dicts = sorted(mgp.cluster_word_distribution[cluster].items(), key=lambda k: k[1], reverse=True)[:num_words]\n",
    "        print(f'Cluster {cluster}: {sort_dicts}')\n"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5db8d6fe",
   "metadata": {},
   "source": [
    " we define the top_words_by_cluster function that prints out the most frequent words that appear in each cluster. It sorts the words in each cluster by their frequency and prints out tuples (of word frequency). The number of words per cluster printed is determined by the num_words parameter."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "fbf027bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "reviews = get_yelp_reviews(yelp_reviews_file)\n",
    "\n",
    " "
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 11,
   "id": "d10722e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "sentences = preprocess(reviews)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 13,
   "id": "d5582de8",
   "metadata": {},
   "outputs": [],
   "source": [
    "vocab = set(word for sentence in sentences for word in sentence)\n",
    "n_terms = len(vocab)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6c90558b",
   "metadata": {},
   "source": [
    "we get a set of all the unique words in the review sentences by turning the list of words into a set, and then we assign the count of these words to the n_terms variable to be used later in the creation of the model."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 14,
   "id": "581608c5",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "In stage 0: transferred 59089 clusters with 25 clusters populated\n",
      "In stage 1: transferred 48396 clusters with 25 clusters populated\n",
      "In stage 2: transferred 40615 clusters with 25 clusters populated\n",
      "In stage 3: transferred 35691 clusters with 25 clusters populated\n",
      "In stage 4: transferred 32481 clusters with 25 clusters populated\n",
      "In stage 5: transferred 30276 clusters with 25 clusters populated\n",
      "In stage 6: transferred 28640 clusters with 25 clusters populated\n",
      "In stage 7: transferred 27141 clusters with 25 clusters populated\n",
      "In stage 8: transferred 26081 clusters with 25 clusters populated\n",
      "In stage 9: transferred 25141 clusters with 24 clusters populated\n",
      "In stage 10: transferred 24460 clusters with 23 clusters populated\n",
      "In stage 11: transferred 24097 clusters with 22 clusters populated\n",
      "In stage 12: transferred 23561 clusters with 22 clusters populated\n",
      "In stage 13: transferred 23288 clusters with 22 clusters populated\n",
      "In stage 14: transferred 23002 clusters with 21 clusters populated\n",
      "In stage 15: transferred 22759 clusters with 20 clusters populated\n",
      "In stage 16: transferred 22447 clusters with 22 clusters populated\n",
      "In stage 17: transferred 22118 clusters with 21 clusters populated\n",
      "In stage 18: transferred 22025 clusters with 19 clusters populated\n",
      "In stage 19: transferred 21940 clusters with 19 clusters populated\n",
      "In stage 20: transferred 21953 clusters with 20 clusters populated\n",
      "In stage 21: transferred 21795 clusters with 20 clusters populated\n",
      "In stage 22: transferred 21550 clusters with 20 clusters populated\n",
      "In stage 23: transferred 21670 clusters with 19 clusters populated\n",
      "In stage 24: transferred 21426 clusters with 19 clusters populated\n",
      "In stage 25: transferred 21307 clusters with 19 clusters populated\n",
      "In stage 26: transferred 21438 clusters with 18 clusters populated\n",
      "In stage 27: transferred 21224 clusters with 18 clusters populated\n",
      "In stage 28: transferred 21070 clusters with 18 clusters populated\n",
      "In stage 29: transferred 20847 clusters with 18 clusters populated\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "[14,\n",
       " 22,\n",
       " 21,\n",
       " 22,\n",
       " 5,\n",
       " 7,\n",
       " 13,\n",
       " 5,\n",
       " 22,\n",
       " 22,\n",
       " 22,\n",
       " 14,\n",
       " 22,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 8,\n",
       " 7,\n",
       " 5,\n",
       " 5,\n",
       " 22,\n",
       " 4,\n",
       " 19,\n",
       " 14,\n",
       " 14,\n",
       " 5,\n",
       " 12,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 8,\n",
       " 8,\n",
       " 7,\n",
       " 14,\n",
       " 19,\n",
       " 8,\n",
       " 14,\n",
       " 14,\n",
       " 22,\n",
       " 5,\n",
       " 14,\n",
       " 7,\n",
       " 0,\n",
       " 21,\n",
       " 7,\n",
       " 7,\n",
       " 12,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 8,\n",
       " 12,\n",
       " 24,\n",
       " 24,\n",
       " 12,\n",
       " 21,\n",
       " 14,\n",
       " 14,\n",
       " 21,\n",
       " 14,\n",
       " 21,\n",
       " 0,\n",
       " 14,\n",
       " 14,\n",
       " 0,\n",
       " 24,\n",
       " 7,\n",
       " 14,\n",
       " 5,\n",
       " 7,\n",
       " 7,\n",
       " 24,\n",
       " 5,\n",
       " 4,\n",
       " 4,\n",
       " 0,\n",
       " 22,\n",
       " 16,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 4,\n",
       " 5,\n",
       " 5,\n",
       " 7,\n",
       " 14,\n",
       " 8,\n",
       " 16,\n",
       " 16,\n",
       " 22,\n",
       " 22,\n",
       " 22,\n",
       " 5,\n",
       " 21,\n",
       " 16,\n",
       " 24,\n",
       " 7,\n",
       " 16,\n",
       " 21,\n",
       " 4,\n",
       " 12,\n",
       " 7,\n",
       " 7,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 7,\n",
       " 21,\n",
       " 4,\n",
       " 4,\n",
       " 8,\n",
       " 21,\n",
       " 24,\n",
       " 14,\n",
       " 21,\n",
       " 4,\n",
       " 14,\n",
       " 19,\n",
       " 19,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 8,\n",
       " 8,\n",
       " 5,\n",
       " 0,\n",
       " 7,\n",
       " 16,\n",
       " 16,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 7,\n",
       " 7,\n",
       " 21,\n",
       " 4,\n",
       " 8,\n",
       " 7,\n",
       " 14,\n",
       " 19,\n",
       " 0,\n",
       " 5,\n",
       " 5,\n",
       " 5,\n",
       " 4,\n",
       " 4,\n",
       " 21,\n",
       " 24,\n",
       " 7,\n",
       " 24,\n",
       " 8,\n",
       " 0,\n",
       " 7,\n",
       " 0,\n",
       " 5,\n",
       " 7,\n",
       " 8,\n",
       " 0,\n",
       " 8,\n",
       " 0,\n",
       " 7,\n",
       " 13,\n",
       " 8,\n",
       " 13,\n",
       " 22,\n",
       " 22,\n",
       " 14,\n",
       " 14,\n",
       " 7,\n",
       " 8,\n",
       " 7,\n",
       " 8,\n",
       " 0,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 22,\n",
       " 5,\n",
       " 4,\n",
       " 4,\n",
       " 13,\n",
       " 14,\n",
       " 8,\n",
       " 4,\n",
       " 8,\n",
       " 8,\n",
       " 7,\n",
       " 14,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 7,\n",
       " 7,\n",
       " 14,\n",
       " 14,\n",
       " 14,\n",
       " 14,\n",
       " 14,\n",
       " 7,\n",
       " 0,\n",
       " 14,\n",
       " 14,\n",
       " 14,\n",
       " 14,\n",
       " 5,\n",
       " 22,\n",
       " 8,\n",
       " 7,\n",
       " 5,\n",
       " 14,\n",
       " 4,\n",
       " 14,\n",
       " 7,\n",
       " 21,\n",
       " 24,\n",
       " 4,\n",
       " 7,\n",
       " 7,\n",
       " 7,\n",
       " 7,\n",
       " 24,\n",
       " 14,\n",
       " 14,\n",
       " 24,\n",
       " 0,\n",
       " 24,\n",
       " 14,\n",
       " 7,\n",
       " 24,\n",
       " 4,\n",
       " 4,\n",
       " 14,\n",
       " 7,\n",
       " 8,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 14,\n",
       " 4,\n",
       " 13,\n",
       " 4,\n",
       " 4,\n",
       " 8,\n",
       " 7,\n",
       " 14,\n",
       " 5,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 14,\n",
       " 4,\n",
       " 21,\n",
       " 14,\n",
       " 12,\n",
       " 7,\n",
       " 7,\n",
       " 4,\n",
       " 16,\n",
       " 19,\n",
       " 4,\n",
       " 4,\n",
       " 19,\n",
       " 4,\n",
       " 5,\n",
       " 14,\n",
       " 14,\n",
       " 19,\n",
       " 5,\n",
       " 22,\n",
       " 22,\n",
       " 22,\n",
       " 22,\n",
       " 5,\n",
       " 7,\n",
       " 0,\n",
       " 0,\n",
       " 24,\n",
       " 8,\n",
       " 4,\n",
       " 8,\n",
       " 5,\n",
       " 5,\n",
       " 4,\n",
       " 14,\n",
       " 5,\n",
       " 8,\n",
       " 4,\n",
       " 8,\n",
       " 5,\n",
       " 22,\n",
       " 7,\n",
       " 13,\n",
       " 5,\n",
       " 19,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 21,\n",
       " 5,\n",
       " 13,\n",
       " 7,\n",
       " 22,\n",
       " 4,\n",
       " 14,\n",
       " 21,\n",
       " 0,\n",
       " 14,\n",
       " 7,\n",
       " 14,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 8,\n",
       " 12,\n",
       " 0,\n",
       " 21,\n",
       " 22,\n",
       " 24,\n",
       " 13,\n",
       " 8,\n",
       " 5,\n",
       " 0,\n",
       " 0,\n",
       " 0,\n",
       " 21,\n",
       " 4,\n",
       " 13,\n",
       " 7,\n",
       " 7,\n",
       " 5,\n",
       " 0,\n",
       " 4,\n",
       " 4,\n",
       " 21,\n",
       " 4,\n",
       " 5,\n",
       " 7,\n",
       " 7,\n",
       " 5,\n",
       " 14,\n",
       " 4,\n",
       " 21,\n",
       " 5,\n",
       " 7,\n",
       " 4,\n",
       " 5,\n",
       " 5,\n",
       " 7,\n",
       " 21,\n",
       " 21,\n",
       " 14,\n",
       " 5,\n",
       " 7,\n",
       " 13,\n",
       " 14,\n",
       " 8,\n",
       " 8,\n",
       " 7,\n",
       " 4,\n",
       " 5,\n",
       " 5,\n",
       " 21,\n",
       " 7,\n",
       " 24,\n",
       " 4,\n",
       " 14,\n",
       " 14,\n",
       " 5,\n",
       " 14,\n",
       " 7,\n",
       " 0,\n",
       " 7,\n",
       " 5,\n",
       " 8,\n",
       " 8,\n",
       " 7,\n",
       " 8,\n",
       " 0,\n",
       " 7,\n",
       " 0,\n",
       " 21,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 8,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 16,\n",
       " 16,\n",
       " 4,\n",
       " 16,\n",
       " 16,\n",
       " 4,\n",
       " 16,\n",
       " 16,\n",
       " 16,\n",
       " 7,\n",
       " 7,\n",
       " 14,\n",
       " 7,\n",
       " 0,\n",
       " 4,\n",
       " 5,\n",
       " 8,\n",
       " 4,\n",
       " 4,\n",
       " 16,\n",
       " 4,\n",
       " 7,\n",
       " 19,\n",
       " 7,\n",
       " 14,\n",
       " 5,\n",
       " 8,\n",
       " 21,\n",
       " 14,\n",
       " 7,\n",
       " 8,\n",
       " 4,\n",
       " 8,\n",
       " 4,\n",
       " 4,\n",
       " 24,\n",
       " 24,\n",
       " 7,\n",
       " 4,\n",
       " 16,\n",
       " 4,\n",
       " 8,\n",
       " 7,\n",
       " 24,\n",
       " 8,\n",
       " 4,\n",
       " 21,\n",
       " 8,\n",
       " 7,\n",
       " 22,\n",
       " 5,\n",
       " 8,\n",
       " 13,\n",
       " 24,\n",
       " 22,\n",
       " 24,\n",
       " 24,\n",
       " 0,\n",
       " 7,\n",
       " 19,\n",
       " 7,\n",
       " 19,\n",
       " 4,\n",
       " 21,\n",
       " 24,\n",
       " 5,\n",
       " 5,\n",
       " 5,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 8,\n",
       " 7,\n",
       " 8,\n",
       " 21,\n",
       " 16,\n",
       " 13,\n",
       " 12,\n",
       " 8,\n",
       " 24,\n",
       " 7,\n",
       " 7,\n",
       " 21,\n",
       " 24,\n",
       " 5,\n",
       " 4,\n",
       " 5,\n",
       " 7,\n",
       " 21,\n",
       " 14,\n",
       " 14,\n",
       " 12,\n",
       " 8,\n",
       " 13,\n",
       " 24,\n",
       " 5,\n",
       " 24,\n",
       " 24,\n",
       " 13,\n",
       " 14,\n",
       " 0,\n",
       " 0,\n",
       " 21,\n",
       " 0,\n",
       " 0,\n",
       " 5,\n",
       " 21,\n",
       " 22,\n",
       " 5,\n",
       " 0,\n",
       " 0,\n",
       " 0,\n",
       " 5,\n",
       " 14,\n",
       " 7,\n",
       " 24,\n",
       " 21,\n",
       " 24,\n",
       " 14,\n",
       " 22,\n",
       " 24,\n",
       " 0,\n",
       " 8,\n",
       " 4,\n",
       " 14,\n",
       " 4,\n",
       " 24,\n",
       " 24,\n",
       " 21,\n",
       " 0,\n",
       " 7,\n",
       " 13,\n",
       " 5,\n",
       " 7,\n",
       " 4,\n",
       " 14,\n",
       " 0,\n",
       " 12,\n",
       " 5,\n",
       " 4,\n",
       " 4,\n",
       " 14,\n",
       " 19,\n",
       " 4,\n",
       " 8,\n",
       " 14,\n",
       " 22,\n",
       " 21,\n",
       " 19,\n",
       " 8,\n",
       " 7,\n",
       " 7,\n",
       " 19,\n",
       " 19,\n",
       " 4,\n",
       " 8,\n",
       " 19,\n",
       " 7,\n",
       " 24,\n",
       " 21,\n",
       " 8,\n",
       " 3,\n",
       " 0,\n",
       " 7,\n",
       " 0,\n",
       " 7,\n",
       " 4,\n",
       " 24,\n",
       " 13,\n",
       " 12,\n",
       " 21,\n",
       " 5,\n",
       " 7,\n",
       " 14,\n",
       " 7,\n",
       " 13,\n",
       " 14,\n",
       " 13,\n",
       " 13,\n",
       " 7,\n",
       " 4,\n",
       " 13,\n",
       " 13,\n",
       " 7,\n",
       " 4,\n",
       " 0,\n",
       " 0,\n",
       " 22,\n",
       " 8,\n",
       " 14,\n",
       " 4,\n",
       " 4,\n",
       " 5,\n",
       " 8,\n",
       " 21,\n",
       " 21,\n",
       " 14,\n",
       " 7,\n",
       " 7,\n",
       " 4,\n",
       " 8,\n",
       " 13,\n",
       " 5,\n",
       " 14,\n",
       " 14,\n",
       " 12,\n",
       " 7,\n",
       " 4,\n",
       " 13,\n",
       " 14,\n",
       " 4,\n",
       " 14,\n",
       " 14,\n",
       " 22,\n",
       " 0,\n",
       " 0,\n",
       " 14,\n",
       " 0,\n",
       " 7,\n",
       " 13,\n",
       " 0,\n",
       " 7,\n",
       " 7,\n",
       " 14,\n",
       " 5,\n",
       " 14,\n",
       " 0,\n",
       " 5,\n",
       " 13,\n",
       " 7,\n",
       " 24,\n",
       " 7,\n",
       " 5,\n",
       " 13,\n",
       " 14,\n",
       " 0,\n",
       " 14,\n",
       " 13,\n",
       " 22,\n",
       " 0,\n",
       " 14,\n",
       " 22,\n",
       " 5,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 22,\n",
       " 24,\n",
       " 14,\n",
       " 7,\n",
       " 0,\n",
       " 8,\n",
       " 0,\n",
       " 0,\n",
       " 7,\n",
       " 7,\n",
       " 14,\n",
       " 21,\n",
       " 24,\n",
       " 22,\n",
       " 14,\n",
       " 5,\n",
       " 14,\n",
       " 22,\n",
       " 7,\n",
       " 7,\n",
       " 19,\n",
       " 21,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 8,\n",
       " 8,\n",
       " 14,\n",
       " 7,\n",
       " 7,\n",
       " 8,\n",
       " 14,\n",
       " 8,\n",
       " 16,\n",
       " 13,\n",
       " 14,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 7,\n",
       " 12,\n",
       " 4,\n",
       " 21,\n",
       " 7,\n",
       " 7,\n",
       " 24,\n",
       " 7,\n",
       " 8,\n",
       " 21,\n",
       " 14,\n",
       " 7,\n",
       " 7,\n",
       " 5,\n",
       " 8,\n",
       " 21,\n",
       " 8,\n",
       " 8,\n",
       " 7,\n",
       " 13,\n",
       " 0,\n",
       " 22,\n",
       " 7,\n",
       " 8,\n",
       " 13,\n",
       " 7,\n",
       " 4,\n",
       " 14,\n",
       " 22,\n",
       " 7,\n",
       " 13,\n",
       " 8,\n",
       " 14,\n",
       " 14,\n",
       " 22,\n",
       " 14,\n",
       " 7,\n",
       " 12,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 5,\n",
       " 24,\n",
       " 21,\n",
       " 5,\n",
       " 13,\n",
       " 5,\n",
       " 7,\n",
       " 14,\n",
       " 0,\n",
       " 13,\n",
       " 14,\n",
       " 4,\n",
       " 7,\n",
       " 16,\n",
       " 22,\n",
       " 16,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 13,\n",
       " 5,\n",
       " 4,\n",
       " 8,\n",
       " 4,\n",
       " 8,\n",
       " 5,\n",
       " 14,\n",
       " 8,\n",
       " 16,\n",
       " 7,\n",
       " 24,\n",
       " 16,\n",
       " 16,\n",
       " 7,\n",
       " 7,\n",
       " 19,\n",
       " 24,\n",
       " 21,\n",
       " 4,\n",
       " 8,\n",
       " 8,\n",
       " 7,\n",
       " 4,\n",
       " 14,\n",
       " 16,\n",
       " 0,\n",
       " 8,\n",
       " 4,\n",
       " 4,\n",
       " 24,\n",
       " 5,\n",
       " 14,\n",
       " 13,\n",
       " 5,\n",
       " 4,\n",
       " 7,\n",
       " 7,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 8,\n",
       " 13,\n",
       " 4,\n",
       " 16,\n",
       " 7,\n",
       " 7,\n",
       " 16,\n",
       " 14,\n",
       " 16,\n",
       " 4,\n",
       " 21,\n",
       " 7,\n",
       " 4,\n",
       " 5,\n",
       " 4,\n",
       " 8,\n",
       " 7,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 19,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 24,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 7,\n",
       " 24,\n",
       " 8,\n",
       " 22,\n",
       " 4,\n",
       " 22,\n",
       " 24,\n",
       " 24,\n",
       " 24,\n",
       " 5,\n",
       " 14,\n",
       " 8,\n",
       " 0,\n",
       " 7,\n",
       " 5,\n",
       " 7,\n",
       " 12,\n",
       " 14,\n",
       " 4,\n",
       " 4,\n",
       " 16,\n",
       " 16,\n",
       " 16,\n",
       " 7,\n",
       " 7,\n",
       " 22,\n",
       " 13,\n",
       " 0,\n",
       " 22,\n",
       " 5,\n",
       " 14,\n",
       " 22,\n",
       " 22,\n",
       " 22,\n",
       " 22,\n",
       " 7,\n",
       " 22,\n",
       " 7,\n",
       " 5,\n",
       " 12,\n",
       " 24,\n",
       " 7,\n",
       " 12,\n",
       " 7,\n",
       " 5,\n",
       " 8,\n",
       " 5,\n",
       " 7,\n",
       " 13,\n",
       " 14,\n",
       " 24,\n",
       " 7,\n",
       " 21,\n",
       " 14,\n",
       " 8,\n",
       " 4,\n",
       " 4,\n",
       " 21,\n",
       " 4,\n",
       " 8,\n",
       " 21,\n",
       " 5,\n",
       " 21,\n",
       " 0,\n",
       " 0,\n",
       " 7,\n",
       " 8,\n",
       " 7,\n",
       " 19,\n",
       " 7,\n",
       " 24,\n",
       " 13,\n",
       " 8,\n",
       " 4,\n",
       " 12,\n",
       " 12,\n",
       " 5,\n",
       " 19,\n",
       " 8,\n",
       " 19,\n",
       " 8,\n",
       " 4,\n",
       " 22,\n",
       " 21,\n",
       " 13,\n",
       " 22,\n",
       " 14,\n",
       " 7,\n",
       " 0,\n",
       " 14,\n",
       " 5,\n",
       " 12,\n",
       " 16,\n",
       " 14,\n",
       " 4,\n",
       " 22,\n",
       " 4,\n",
       " 7,\n",
       " 24,\n",
       " 24,\n",
       " 5,\n",
       " 5,\n",
       " 22,\n",
       " 0,\n",
       " 22,\n",
       " 22,\n",
       " 14,\n",
       " 0,\n",
       " 22,\n",
       " 0,\n",
       " 7,\n",
       " 22,\n",
       " 5,\n",
       " 22,\n",
       " 22,\n",
       " 7,\n",
       " 5,\n",
       " 5,\n",
       " 5,\n",
       " 8,\n",
       " 4,\n",
       " 5,\n",
       " 5,\n",
       " 21,\n",
       " 21,\n",
       " 14,\n",
       " 7,\n",
       " 7,\n",
       " 14,\n",
       " 14,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 19,\n",
       " 19,\n",
       " 4,\n",
       " 21,\n",
       " 21,\n",
       " 8,\n",
       " 21,\n",
       " 13,\n",
       " 16,\n",
       " 16,\n",
       " 14,\n",
       " 16,\n",
       " 7,\n",
       " 8,\n",
       " 14,\n",
       " 8,\n",
       " 21,\n",
       " 13,\n",
       " 24,\n",
       " 7,\n",
       " 4,\n",
       " 16,\n",
       " 8,\n",
       " 7,\n",
       " 4,\n",
       " 4,\n",
       " 4,\n",
       " 14,\n",
       " ...]"
      ]
     },
     "execution_count": 14,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "#Fit the Model\n",
    "mgp = MovieGroupProcess(K=25, alpha=0.1, beta=0.1, n_iters=30)\n",
    "mgp.fit(sentences, n_terms)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "60592591",
   "metadata": {},
   "source": [
    "we create the GSDMM model. The K parameter is the upper bound on the number of clusters, as the algorithm determines the number of clusters less than or equal to this number. The alpha parameter controls the probability that a new cluster will be created, and the beta parameter defines how new text is clustered. If the value of beta is closer to 0, then text will be clustered more according to similarity, while if it is closer to 1,\n",
    "the clustering will be more based on the frequency of texts. The n_iters parameter determines the number of passes the algorithm makes through the corpus."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 15,
   "id": "5f22bdc0",
   "metadata": {},
   "outputs": [],
   "source": [
    "'''we get the count of documents by topic and then create a list of the 15 most populous topics. \n",
    "We then use this list in step 9 to get the 10 most frequent words in each cluster.'''  \n",
    "doc_count = np.array(mgp.cluster_doc_count)\n",
    "top_clusters = doc_count.argsort()[-18:][::-1]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "5f9efd2e",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "Cluster 4: [('chicken', 985), ('ordered', 872), ('cheese', 780), ('sauce', 765), ('delicious', 649), ('salad', 578), ('fries', 540), ('fresh', 516), ('shrimp', 457), ('fried', 456), ('sandwich', 403), ('burger', 394), ('side', 367), ('cooked', 337), ('bread', 330), ('pork', 324), ('order', 322), ('meat', 321), ('pizza', 307), ('hot', 304)]\n",
      "Cluster 14: [('order', 671), ('table', 539), ('wait', 498), ('people', 388), ('service', 374), ('asked', 328), ('bar', 310), ('told', 295), ('long', 282), ('waitress', 267), ('night', 264), ('server', 256), ('seated', 244), ('ordered', 241), ('restaurant', 235), ('waiting', 233), ('drink', 228), ('drinks', 226), ('busy', 223), ('menu', 217)]\n",
      "Cluster 7: [('service', 776), ('experience', 365), ('restaurant', 364), ('worth', 364), ('stars', 316), ('lunch', 279), ('eat', 266), ('location', 249), ('wait', 244), ('breakfast', 210), ('night', 205), ('5', 203), ('spot', 198), ('visit', 191), ('hotel', 190), ('star', 189), ('area', 185), ('find', 180), ('places', 174), ('people', 174)]\n",
      "Cluster 8: [('menu', 437), ('beer', 373), ('selection', 362), ('prices', 285), ('price', 264), ('service', 251), ('wine', 213), ('drinks', 203), ('delicious', 198), ('fresh', 179), ('bar', 161), ('options', 155), ('beers', 152), ('reasonable', 150), ('quality', 148), ('priced', 144), ('items', 141), ('portions', 136), ('small', 128), ('variety', 119)]\n",
      "Cluster 24: [('room', 438), ('clean', 300), ('area', 266), ('bar', 231), ('seating', 195), ('atmosphere', 183), ('small', 178), ('inside', 168), ('restaurant', 156), ('music', 151), ('tables', 143), ('decor', 136), ('people', 133), ('floor', 132), ('space', 130), ('hotel', 128), ('rooms', 127), ('beautiful', 124), ('comfortable', 122), ('cool', 107)]\n",
      "Cluster 5: [('service', 1052), ('friendly', 1006), ('staff', 885), ('helpful', 303), ('attentive', 221), ('customer', 172), ('atmosphere', 167), ('clean', 155), ('server', 150), ('people', 131), ('fast', 122), ('knowledgeable', 119), ('experience', 118), ('professional', 110), ('owner', 105), ('work', 102), ('waitress', 97), ('pleasant', 94), ('wait', 90), ('restaurant', 80)]\n",
      "Cluster 0: [('car', 226), ('told', 199), ('called', 185), ('phone', 157), ('asked', 142), ('call', 136), ('service', 126), ('front', 126), ('order', 119), ('manager', 115), ('room', 109), ('work', 106), ('desk', 100), ('price', 99), ('check', 95), ('card', 92), ('needed', 91), ('appointment', 81), ('wanted', 73), ('left', 70)]\n",
      "Cluster 21: [('restaurant', 257), ('favorite', 168), ('area', 167), ('places', 120), ('pizza', 116), ('restaurants', 114), ('mexican', 110), ('service', 106), ('authentic', 102), ('find', 101), ('city', 96), ('town', 95), ('experience', 93), ('local', 90), ('eat', 89), ('sushi', 88), ('highly', 83), ('italian', 77), ('family', 76), ('spot', 75)]\n",
      "Cluster 22: [('hair', 220), ('job', 108), ('cut', 100), ('experience', 88), ('work', 85), ('nails', 80), ('salon', 78), ('massage', 67), ('nail', 66), ('wanted', 65), ('service', 61), ('needed', 60), ('care', 55), ('gel', 48), ('dr.', 46), ('color', 45), ('left', 44), ('told', 44), ('pedicure', 43), ('wedding', 42)]\n",
      "Cluster 16: [('cream', 265), ('ice', 236), ('chocolate', 208), ('coffee', 161), ('cake', 141), ('delicious', 136), ('sweet', 131), ('tea', 116), ('ordered', 103), ('flavor', 84), ('dessert', 81), ('tasted', 81), ('bread', 71), ('banana', 71), ('fresh', 69), ('butter', 68), ('pudding', 65), ('sugar', 64), ('taste', 63), ('milk', 63)]\n",
      "Cluster 13: [('night', 182), ('music', 125), ('dinner', 119), ('live', 91), ('experience', 86), ('lunch', 86), ('friends', 78), ('birthday', 76), ('sunday', 76), ('people', 74), ('saturday', 69), ('kids', 62), ('brunch', 61), ('bar', 59), ('game', 59), ('fun', 58), ('weekend', 58), ('highly', 49), ('town', 49), ('family', 48)]\n",
      "Cluster 12: [('location', 156), ('parking', 130), ('street', 121), ('city', 88), ('park', 80), ('store', 70), ('located', 63), ('walk', 62), ('restaurant', 59), ('mall', 58), ('walking', 53), ('area', 52), ('center', 52), ('find', 49), ('hotel', 48), ('bar', 46), ('shopping', 45), ('people', 42), ('strip', 41), ('tour', 39)]\n",
      "Cluster 19: [('chicken', 174), ('rice', 74), ('spicy', 70), ('ordered', 67), ('beef', 65), ('soup', 64), ('dish', 61), ('noodles', 61), ('rolls', 61), ('flavor', 60), ('fried', 55), ('delicious', 52), ('roll', 52), ('thai', 51), ('curry', 50), ('meat', 49), ('dishes', 47), ('favorite', 44), ('sushi', 42), ('sauce', 41)]\n",
      "Cluster 3: [('custom', 3), ('vintage', 3), ('fast', 3), ('smoked', 3), ('branches', 3), ('plantation', 2), ('laura', 2), ('elimination', 2), ('legumes', 2), ('copper', 2), ('plantations', 2), ('bronx', 2), ('amazed', 2), ('bottle', 2), ('benefits', 2), ('brass', 2), ('olin', 2), ('silver', 2), ('informative', 2), ('tour', 1)]\n",
      "Cluster 10: [('kanella', 2), ('echo', 2), ('speckled', 1), ('wires', 1), ('designed', 1), ('papers', 1), ('artistically', 1), ('striking', 1), ('elements', 1), ('luxe', 1), ('masterpieces', 1), ('creatively', 1), ('fabrics', 1), ('special.ok', 1), ('uniquely', 1), ('bons', 1), ('les', 1), ('rouler', 1), ('temps', 1), ('laissez', 1)]\n",
      "Cluster 11: [('yelp', 3), ('orleans', 3), ('visiting', 3), ('includes', 3), ('recommendations', 3), ('share', 3), ('message', 3), ('itinerary', 3), ('st.louis', 1), ('......', 1), ('sayin', 1), ('jus', 1), ('hendricks', 1), ('treatment', 1), ('post-assembly', 1), ('elbow', 1), ('grease', 1), ('14', 1), ('dudes', 1), ('tchoup', 1)]\n",
      "Cluster 23: [('robert', 2), ('indiana', 2), ('sculpture', 2), ('iconic', 2), ('pigs', 2), ('-especially', 1), ('carbaholics', 1), ('hills', 1), ('dine', 1), ('opportunity', 1), ('foot', 1), ('stay-cation', 1), ('fiamme', 1), ('jumped', 1), ('drawings', 1), ('britain', 1), ('painted', 1), ('ill-placed', 1), ('include', 1), ('art', 1)]\n",
      "Cluster 1: [('hong', 2), ('kong', 2), ('mandarin', 2), ('matt', 2), ('cantonese', 2), ('akira', 1), ('exception', 1), ('singer', 1), ('version', 1), ('lai', 1), ('andy', 1), ('kwok', 1), ('aaron', 1), (\"1980's-\", 1), ('formed', 1), ('cheung', 1), ('fans', 1), ('popular', 1), ('mid', 1), ('jacky', 1)]\n"
     ]
    }
   ],
   "source": [
    "top_words_by_cluster(mgp, top_clusters, 20)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "5f38a6f0",
   "metadata": {},
   "source": [
    "The results of the clustering make sense for many of the clusters. In the preceding results, cluster 6 is about food, clusters 4 and 9 relate to the service, cluster 5 is about the selection available, cluster 24 concerns the atmosphere, cluster 3 is about dessert, and cluster 18 is about hair and nail salons."
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}