Yelp_NLP_Topic_Modeling.ipynb

{
  "nbformat": 4,
  "nbformat_minor": 0,
  "metadata": {
    "colab": {
      "provenance": [],
      "collapsed_sections": [],
      "authorship_tag": "ABX9TyPYi9UI5czDcNW1S54jPhZH",
      "include_colab_link": true
    },
    "kernelspec": {
      "name": "python3",
      "display_name": "Python 3"
    },
    "language_info": {
      "name": "python"
    }
  },
  "cells": [
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "view-in-github",
        "colab_type": "text"
      },
      "source": [
        "<a href=\"https://colab.research.google.com/github/EAName/yelp-reviews/blob/main/Yelp_NLP_Topic_Modeling.ipynb\" target=\"_parent\"><img src=\"https://colab.research.google.com/assets/colab-badge.svg\" alt=\"Open In Colab\"/></a>"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 1,
      "metadata": {
        "id": "92Dg33FmuxQs"
      },
      "outputs": [],
      "source": [
        "import pandas as pd\n",
        "import numpy as np\n",
        "import os\n",
        "import random\n",
        "import numpy as np\n",
        "import pickle \n",
        "import re,string\n",
        "\n",
        "import nltk\n",
        "from nltk.corpus import stopwords\n",
        "from nltk.stem import PorterStemmer\n",
        "from nltk.stem import WordNetLemmatizer\n",
        "from nltk.tokenize import word_tokenize \n",
        "\n",
        "from sklearn.feature_extraction.text import TfidfVectorizer\n",
        "from sklearn.feature_extraction.text import CountVectorizer\n",
        "from sklearn.cluster import KMeans, AgglomerativeClustering\n",
        "from sklearn.metrics.pairwise import cosine_similarity\n",
        "from sklearn.decomposition import TruncatedSVD, LatentDirichletAllocation\n",
        "from sklearn.svm import SVC\n",
        "from sklearn.preprocessing import LabelEncoder, OneHotEncoder\n",
        "from sklearn.ensemble import RandomForestClassifier\n",
        "from sklearn.metrics import roc_auc_score, accuracy_score, confusion_matrix, silhouette_score\n",
        "from sklearn.metrics import precision_score, recall_score, f1_score, classification_report\n",
        "from sklearn.metrics import mean_squared_error as MSE\n",
        "from sklearn.metrics import mean_absolute_error\n",
        "from sklearn.model_selection import cross_val_score, StratifiedKFold\n",
        "from sklearn.tree import DecisionTreeClassifier\n",
        "from sklearn.naive_bayes import MultinomialNB\n",
        "from sklearn.linear_model import LogisticRegression\n",
        "from sklearn.ensemble import RandomForestRegressor\n",
        "from sklearn.metrics import accuracy_score\n",
        "from sklearn.model_selection import train_test_split, KFold\n",
        "from sklearn.manifold import MDS\n",
        "from sklearn.metrics.pairwise import cosine_similarity\n",
        "from sklearn.ensemble import GradientBoostingClassifier\n",
        "from sklearn.svm import LinearSVC\n",
        "\n",
        "\n",
        "\n",
        "from sklearn.model_selection import train_test_split \n",
        "from sklearn.model_selection import GridSearchCV \n",
        "from sklearn.model_selection import cross_val_score \n",
        "from sklearn.model_selection import StratifiedKFold\n",
        "from sklearn.model_selection import KFold \n",
        "from sklearn.model_selection import cross_val_predict\n",
        "\n",
        "\n",
        "import seaborn as sns\n",
        "import matplotlib.pyplot as plt\n",
        "\n",
        "from gensim.models import Word2Vec,LdaMulticore, TfidfModel, CoherenceModel\n",
        "from gensim import corpora\n",
        "from gensim import similarities\n",
        "from gensim.models.doc2vec import Doc2Vec, TaggedDocument\n",
        "from gensim.models import LsiModel,LdaModel\n",
        "\n",
        "from keras.preprocessing.text import Tokenizer\n",
        "#from keras.preprocessing.sequence import pad_sequences\n",
        "from tqdm import tqdm\n",
        "from keras.models import Sequential\n",
        "from keras.layers import Embedding, LSTM, Dense, SpatialDropout1D, GRU\n",
        "from keras.callbacks import EarlyStopping\n",
        "from keras.initializers import Constant\n",
        "from sklearn.model_selection import train_test_split\n",
        "from tensorflow.keras.optimizers import Adam, Nadam, RMSprop\n",
        "from keras import Input\n",
        "\n",
        "from IPython.display import display"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "nltk.download('stopwords')\n",
        "nltk.download('wordnet')\n",
        "nltk.download('omw-1.4')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "VGHly-zIpris",
        "outputId": "277525cc-6e20-4025-d655-beef80f06f88"
      },
      "execution_count": 2,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "[nltk_data] Downloading package stopwords to /root/nltk_data...\n",
            "[nltk_data]   Package stopwords is already up-to-date!\n",
            "[nltk_data] Downloading package wordnet to /root/nltk_data...\n",
            "[nltk_data]   Package wordnet is already up-to-date!\n",
            "[nltk_data] Downloading package omw-1.4 to /root/nltk_data...\n",
            "[nltk_data]   Package omw-1.4 is already up-to-date!\n"
          ]
        },
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "True"
            ]
          },
          "metadata": {},
          "execution_count": 2
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "import warnings\n",
        "warnings.filterwarnings('ignore')"
      ],
      "metadata": {
        "id": "ovbC3RlUpxKS"
      },
      "execution_count": 3,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "from google.colab import drive\n",
        "drive.mount('/content/gdrive')"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "xqTk3PNGvHH2",
        "outputId": "09bf078a-ede2-4ee7-94e8-67e770e80356"
      },
      "execution_count": 4,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Drive already mounted at /content/gdrive; to attempt to forcibly remount, call drive.mount(\"/content/gdrive\", force_remount=True).\n"
          ]
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df = pd.read_csv('/content/gdrive/MyDrive/Data/yelp_reviews_clean_version4_st_louis.csv')"
      ],
      "metadata": {
        "id": "ofrho4sfu5nc"
      },
      "execution_count": 5,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.shape"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "MfQeqg9yvQnR",
        "outputId": "8dd77b5e-4c30-4800-ba96-4d7de7557eae"
      },
      "execution_count": 6,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "(292376, 5)"
            ]
          },
          "metadata": {},
          "execution_count": 6
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,6))\n",
        "ax = sns.countplot(df['stars'])\n",
        "plt.title('Distribution of rating');\n",
        "plt.title(\"Star Rating Distribution\")\n",
        "plt.ylabel('# of businesses', fontsize=12)\n",
        "plt.xlabel('Star Ratings ', fontsize=12)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "Ab2dpm8VYBUR",
        "outputId": "62763bfe-8a19-46ae-c28f-73763c0ab220"
      },
      "execution_count": 7,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'Star Ratings ')"
            ]
          },
          "metadata": {},
          "execution_count": 7
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# sample 10,000 reviews\n",
        "df = df.sample(n=10000, random_state=1)"
      ],
      "metadata": {
        "id": "19JslLhFXt5g"
      },
      "execution_count": 8,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "plt.figure(figsize=(10,6))\n",
        "ax = sns.countplot(df['stars'])\n",
        "plt.title('Distribution of rating');\n",
        "plt.title(\"Star Rating Distribution\")\n",
        "plt.ylabel('# of businesses', fontsize=12)\n",
        "plt.xlabel('Star Ratings ', fontsize=12)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 424
        },
        "id": "cRiti7FkYQbA",
        "outputId": "c7db1ef1-06ae-4e0c-f322-e04555470dae"
      },
      "execution_count": 9,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "Text(0.5, 0, 'Star Ratings ')"
            ]
          },
          "metadata": {},
          "execution_count": 9
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 720x432 with 1 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Ede7p75luxQy"
      },
      "source": [
        "**Extract year**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 10,
      "metadata": {
        "id": "SggCnX1zuxQy"
      },
      "outputs": [],
      "source": [
        "df['review_year'] = df['date'].apply(lambda x:int(x[:4]))"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "-_4Eohp3uxQy"
      },
      "source": [
        "**Look at distribution of reviews by year**"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "df.review_year.value_counts(ascending= True)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "id": "8BiT5Omkvh9q",
        "outputId": "58e7487a-cdd3-41a8-aee8-1e4187015527"
      },
      "execution_count": 11,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "2006      11\n",
              "2007      24\n",
              "2022      32\n",
              "2008     101\n",
              "2009     112\n",
              "2010     308\n",
              "2011     456\n",
              "2012     531\n",
              "2013     616\n",
              "2020     646\n",
              "2021     746\n",
              "2014     779\n",
              "2015     998\n",
              "2016    1079\n",
              "2017    1127\n",
              "2018    1181\n",
              "2019    1253\n",
              "Name: review_year, dtype: int64"
            ]
          },
          "metadata": {},
          "execution_count": 11
        }
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "Du6o1gITuxQy"
      },
      "source": [
        "**Create 1-0 sentiment flag**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 12,
      "metadata": {
        "id": "zH9VbEWHuxQy"
      },
      "outputs": [],
      "source": [
        "df['sentiment'] = df['stars'].apply(lambda x: 1 if x >= 4 else 0)"
      ]
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "T7yBmnQguxQz"
      },
      "source": [
        "**Drop reviews older than 2011**"
      ]
    },
    {
      "cell_type": "code",
      "execution_count": 13,
      "metadata": {
        "id": "v3QeIOP6uxQz"
      },
      "outputs": [],
      "source": [
        "#df = df[(df['review_year'] > 2011)]"
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#df.shape"
      ],
      "metadata": {
        "id": "z9TQ9u8L4QfF"
      },
      "execution_count": 14,
      "outputs": []
    },
    {
      "cell_type": "markdown",
      "metadata": {
        "id": "f2OVxMlpuxQz"
      },
      "source": [
        "## Text Preprocessing "
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "def clean_doc(doc): \n",
        "    #split review into individual words\n",
        "    tokens=doc.split()\n",
        "    re_punc = re.compile('[%s]' % re.escape(string.punctuation))\n",
        "    # remove punctuation from each word\n",
        "    tokens = [re_punc.sub('', w) for w in tokens]\n",
        "\n",
        "    #word lemmzatization    \n",
        "    lemm = WordNetLemmatizer()\n",
        "    tokens=[lemm.lemmatize(word) for word in tokens]\n",
        "\n",
        "    # remove remaining tokens that are not alphabetic\n",
        "    tokens = [word for word in tokens if word.isalpha()]\n",
        "    # filter out short tokens\n",
        "    tokens = [word for word in tokens if len(word) > 3]\n",
        "    #lowercase all words\n",
        "    tokens = [word.lower() for word in tokens]\n",
        "    # filter out stop words\n",
        "    stop_words = set(stopwords.words('english'))\n",
        "    tokens = [w for w in tokens if not w in stop_words]     \n",
        "\n",
        "    # filter out common words\n",
        "    #tokens = [w for w in tokens if not w in ['movie', 'movies', 'films']]      \n",
        "    # # word stemming    \n",
        "    # ps=PorterStemmer()\n",
        "    # tokens=[ps.stem(word) for word in tokens]\n",
        "\n",
        "    return tokens\n",
        "\n",
        "def final_processed_text(doc):\n",
        "    #this is a function to join the processed text back\n",
        "    ' '.join(doc)\n",
        "    return doc"
      ],
      "metadata": {
        "id": "9aA2s_9J1hao"
      },
      "execution_count": 15,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df['cleaned_text'] = df['text'].apply(lambda x: clean_doc(x))\n",
        "\n",
        "final_processed_text = [' '.join(x) for x in df['cleaned_text'].tolist()]\n",
        "df['final_processed_text'] = final_processed_text     #add to dataframe"
      ],
      "metadata": {
        "id": "cFEYh7_b1hc0"
      },
      "execution_count": 16,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df['text length'] = df['text'].apply(len)"
      ],
      "metadata": {
        "id": "3jczJFSds5vH"
      },
      "execution_count": 17,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "final_processed_text[0]"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 87
        },
        "id": "KOC1OSzkbri6",
        "outputId": "15a07bc3-4484-4c46-b21a-fd65de1c82f4"
      },
      "execution_count": 18,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "'came around third person party wasnt supposed arrive around seated promptly even without full party dont busy ordered risotto ball couldnt wait third person arrive digging gooey little bite deliciousness order offered fresh sliced bread dont deny thats darn good bread everyone arrived ordered pasta suggestion waitress helpful knowledgeable menu also asked mixed soda drink wasnt cherry limeade said bartender could something ended previous month specialty drink raspberry black pepper meal complete enjoyed tasting ending meal gelato well back soon phenomenal looking pizza update came back pizza gelato baseball game enjoyed especially foodie'"
            ],
            "application/vnd.google.colaboratory.intrinsic+json": {
              "type": "string"
            }
          },
          "metadata": {},
          "execution_count": 18
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "#df.to_csv('/content/gdrive/MyDrive/data/yelp_reviews_pre_processed_v2.csv', index = False)"
      ],
      "metadata": {
        "id": "m2gkVi83txjl"
      },
      "execution_count": 19,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "#df = pd.read_csv('/content/gdrive/MyDrive/data/yelp_reviews_pre_processed_v2.csv')"
      ],
      "metadata": {
        "id": "UdiIQGmkamjR"
      },
      "execution_count": 20,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "df.head()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 441
        },
        "id": "OmWMPf4vt4M9",
        "outputId": "913e5758-1cab-4210-c71d-4f1e594414cb"
      },
      "execution_count": 21,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "                       user_id             business_id                 date  \\\n",
              "34977   J77sLuzhywuyDI2ro0ySkQ  YSgcojsb2kWHXNLWaGCuBA  2016-07-26 19:43:56   \n",
              "225241  Nzy7auzEFEN5wthmL9a00w  jME5VhjMZvNKsQJwsPKL7Q  2016-08-22 16:50:49   \n",
              "202828  VZGTPhVezGAIxsYgrhCLyw  s7JxSEtIumoFod2Jdj0_-g  2018-11-22 12:00:07   \n",
              "107110  o15SFerEo23dMSiY_P2xEw  oC91lprbbfzw7DOYiYln7Q  2014-11-16 18:05:18   \n",
              "52350   gFttJC8-aX5bXcUuqyH5CA  9llupBLdHKFzfawIk6I74w  2013-08-30 01:16:57   \n",
              "\n",
              "        stars                                               text  review_year  \\\n",
              "34977     5.0  My mom and I came in around 5:15 and the third...         2016   \n",
              "225241    4.0  Great place for lunch. Inside and Outside seat...         2016   \n",
              "202828    1.0  Horrible experience! Owner is completely rude ...         2018   \n",
              "107110    4.0  Didn't realize I hadn't written review on this...         2014   \n",
              "52350     4.0  I have been here like 5 times and everytime I ...         2013   \n",
              "\n",
              "        sentiment                                       cleaned_text  \\\n",
              "34977           1  [came, around, third, person, party, wasnt, su...   \n",
              "225241          1  [great, place, lunch, inside, outside, seating...   \n",
              "202828          0  [horrible, experience, owner, completely, rude...   \n",
              "107110          1  [didnt, realize, hadnt, written, review, place...   \n",
              "52350           1  [like, time, everytime, like, coming, back, re...   \n",
              "\n",
              "                                     final_processed_text  text length  \n",
              "34977   came around third person party wasnt supposed ...         1065  \n",
              "225241  great place lunch inside outside seating food ...          172  \n",
              "202828  horrible experience owner completely rude unpr...          193  \n",
              "107110  didnt realize hadnt written review place good ...          893  \n",
              "52350   like time everytime like coming back restauran...         1554  "
            ],
            "text/html": [
              "\n",
              "  <div id=\"df-a4638862-8c36-442b-9326-271090100338\">\n",
              "    <div class=\"colab-df-container\">\n",
              "      <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>user_id</th>\n",
              "      <th>business_id</th>\n",
              "      <th>date</th>\n",
              "      <th>stars</th>\n",
              "      <th>text</th>\n",
              "      <th>review_year</th>\n",
              "      <th>sentiment</th>\n",
              "      <th>cleaned_text</th>\n",
              "      <th>final_processed_text</th>\n",
              "      <th>text length</th>\n",
              "    </tr>\n",
              "  </thead>\n",
              "  <tbody>\n",
              "    <tr>\n",
              "      <th>34977</th>\n",
              "      <td>J77sLuzhywuyDI2ro0ySkQ</td>\n",
              "      <td>YSgcojsb2kWHXNLWaGCuBA</td>\n",
              "      <td>2016-07-26 19:43:56</td>\n",
              "      <td>5.0</td>\n",
              "      <td>My mom and I came in around 5:15 and the third...</td>\n",
              "      <td>2016</td>\n",
              "      <td>1</td>\n",
              "      <td>[came, around, third, person, party, wasnt, su...</td>\n",
              "      <td>came around third person party wasnt supposed ...</td>\n",
              "      <td>1065</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>225241</th>\n",
              "      <td>Nzy7auzEFEN5wthmL9a00w</td>\n",
              "      <td>jME5VhjMZvNKsQJwsPKL7Q</td>\n",
              "      <td>2016-08-22 16:50:49</td>\n",
              "      <td>4.0</td>\n",
              "      <td>Great place for lunch. Inside and Outside seat...</td>\n",
              "      <td>2016</td>\n",
              "      <td>1</td>\n",
              "      <td>[great, place, lunch, inside, outside, seating...</td>\n",
              "      <td>great place lunch inside outside seating food ...</td>\n",
              "      <td>172</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>202828</th>\n",
              "      <td>VZGTPhVezGAIxsYgrhCLyw</td>\n",
              "      <td>s7JxSEtIumoFod2Jdj0_-g</td>\n",
              "      <td>2018-11-22 12:00:07</td>\n",
              "      <td>1.0</td>\n",
              "      <td>Horrible experience! Owner is completely rude ...</td>\n",
              "      <td>2018</td>\n",
              "      <td>0</td>\n",
              "      <td>[horrible, experience, owner, completely, rude...</td>\n",
              "      <td>horrible experience owner completely rude unpr...</td>\n",
              "      <td>193</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>107110</th>\n",
              "      <td>o15SFerEo23dMSiY_P2xEw</td>\n",
              "      <td>oC91lprbbfzw7DOYiYln7Q</td>\n",
              "      <td>2014-11-16 18:05:18</td>\n",
              "      <td>4.0</td>\n",
              "      <td>Didn't realize I hadn't written review on this...</td>\n",
              "      <td>2014</td>\n",
              "      <td>1</td>\n",
              "      <td>[didnt, realize, hadnt, written, review, place...</td>\n",
              "      <td>didnt realize hadnt written review place good ...</td>\n",
              "      <td>893</td>\n",
              "    </tr>\n",
              "    <tr>\n",
              "      <th>52350</th>\n",
              "      <td>gFttJC8-aX5bXcUuqyH5CA</td>\n",
              "      <td>9llupBLdHKFzfawIk6I74w</td>\n",
              "      <td>2013-08-30 01:16:57</td>\n",
              "      <td>4.0</td>\n",
              "      <td>I have been here like 5 times and everytime I ...</td>\n",
              "      <td>2013</td>\n",
              "      <td>1</td>\n",
              "      <td>[like, time, everytime, like, coming, back, re...</td>\n",
              "      <td>like time everytime like coming back restauran...</td>\n",
              "      <td>1554</td>\n",
              "    </tr>\n",
              "  </tbody>\n",
              "</table>\n",
              "</div>\n",
              "      <button class=\"colab-df-convert\" onclick=\"convertToInteractive('df-a4638862-8c36-442b-9326-271090100338')\"\n",
              "              title=\"Convert this dataframe to an interactive table.\"\n",
              "              style=\"display:none;\">\n",
              "        \n",
              "  <svg xmlns=\"http://www.w3.org/2000/svg\" height=\"24px\"viewBox=\"0 0 24 24\"\n",
              "       width=\"24px\">\n",
              "    <path d=\"M0 0h24v24H0V0z\" fill=\"none\"/>\n",
              "    <path d=\"M18.56 5.44l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94zm-11 1L8.5 8.5l.94-2.06 2.06-.94-2.06-.94L8.5 2.5l-.94 2.06-2.06.94zm10 10l.94 2.06.94-2.06 2.06-.94-2.06-.94-.94-2.06-.94 2.06-2.06.94z\"/><path d=\"M17.41 7.96l-1.37-1.37c-.4-.4-.92-.59-1.43-.59-.52 0-1.04.2-1.43.59L10.3 9.45l-7.72 7.72c-.78.78-.78 2.05 0 2.83L4 21.41c.39.39.9.59 1.41.59.51 0 1.02-.2 1.41-.59l7.78-7.78 2.81-2.81c.8-.78.8-2.07 0-2.86zM5.41 20L4 18.59l7.72-7.72 1.47 1.35L5.41 20z\"/>\n",
              "  </svg>\n",
              "      </button>\n",
              "      \n",
              "  <style>\n",
              "    .colab-df-container {\n",
              "      display:flex;\n",
              "      flex-wrap:wrap;\n",
              "      gap: 12px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert {\n",
              "      background-color: #E8F0FE;\n",
              "      border: none;\n",
              "      border-radius: 50%;\n",
              "      cursor: pointer;\n",
              "      display: none;\n",
              "      fill: #1967D2;\n",
              "      height: 32px;\n",
              "      padding: 0 0 0 0;\n",
              "      width: 32px;\n",
              "    }\n",
              "\n",
              "    .colab-df-convert:hover {\n",
              "      background-color: #E2EBFA;\n",
              "      box-shadow: 0px 1px 2px rgba(60, 64, 67, 0.3), 0px 1px 3px 1px rgba(60, 64, 67, 0.15);\n",
              "      fill: #174EA6;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert {\n",
              "      background-color: #3B4455;\n",
              "      fill: #D2E3FC;\n",
              "    }\n",
              "\n",
              "    [theme=dark] .colab-df-convert:hover {\n",
              "      background-color: #434B5C;\n",
              "      box-shadow: 0px 1px 3px 1px rgba(0, 0, 0, 0.15);\n",
              "      filter: drop-shadow(0px 1px 2px rgba(0, 0, 0, 0.3));\n",
              "      fill: #FFFFFF;\n",
              "    }\n",
              "  </style>\n",
              "\n",
              "      <script>\n",
              "        const buttonEl =\n",
              "          document.querySelector('#df-a4638862-8c36-442b-9326-271090100338 button.colab-df-convert');\n",
              "        buttonEl.style.display =\n",
              "          google.colab.kernel.accessAllowed ? 'block' : 'none';\n",
              "\n",
              "        async function convertToInteractive(key) {\n",
              "          const element = document.querySelector('#df-a4638862-8c36-442b-9326-271090100338');\n",
              "          const dataTable =\n",
              "            await google.colab.kernel.invokeFunction('convertToInteractive',\n",
              "                                                     [key], {});\n",
              "          if (!dataTable) return;\n",
              "\n",
              "          const docLinkHtml = 'Like what you see? Visit the ' +\n",
              "            '<a target=\"_blank\" href=https://colab.research.google.com/notebooks/data_table.ipynb>data table notebook</a>'\n",
              "            + ' to learn more about interactive tables.';\n",
              "          element.innerHTML = '';\n",
              "          dataTable['output_type'] = 'display_data';\n",
              "          await google.colab.output.renderOutput(dataTable, element);\n",
              "          const docLink = document.createElement('div');\n",
              "          docLink.innerHTML = docLinkHtml;\n",
              "          element.appendChild(docLink);\n",
              "        }\n",
              "      </script>\n",
              "    </div>\n",
              "  </div>\n",
              "  "
            ]
          },
          "metadata": {},
          "execution_count": 21
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "# Lets visualise the review text and stars for a business\n",
        "g = sns.FacetGrid(data=df, col='stars')\n",
        "g.map(plt.hist, 'text length', bins=50)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 242
        },
        "id": "dRBhqCZ_s9LJ",
        "outputId": "ec6f8eae-11ec-4d6d-b76b-fd1f41f7be1c"
      },
      "execution_count": 22,
      "outputs": [
        {
          "output_type": "execute_result",
          "data": {
            "text/plain": [
              "<seaborn.axisgrid.FacetGrid at 0x7f94bb996090>"
            ]
          },
          "metadata": {},
          "execution_count": 22
        },
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 1080x216 with 5 Axes>"
            ],
            "image/png": "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\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "###K-Mean Clustering"
      ],
      "metadata": {
        "id": "w0Apfj--SykF"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "def tfidf(corpus, titles, ngram_range = (1,1)):\n",
        "    #this is a function to created the tfidf matrix\n",
        "    Tfidf=TfidfVectorizer(ngram_range= ngram_range, max_features = 2000)\n",
        "\n",
        "    #fit the vectorizer using final processed documents.  The vectorizer requires the \n",
        "    #stiched back together document.\n",
        "\n",
        "    TFIDF_matrix=Tfidf.fit_transform(corpus)     \n",
        "\n",
        "    #creating datafram from TFIDF Matrix\n",
        "    words = Tfidf.get_feature_names()\n",
        "    matrix=pd.DataFrame(TFIDF_matrix.toarray(), columns=Tfidf.get_feature_names(), index=titles)\n",
        "    return matrix, words"
      ],
      "metadata": {
        "id": "YSOtJ4zdSz_-"
      },
      "execution_count": 66,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def k_means(titles, tfidf_matrix, k=3):\n",
        "    \n",
        "    #this is a function to generate the k-means output using the tfidf matrix.  Inputs \n",
        "    #to the function include: titles of text, processed text, and desired k value. \n",
        "    #Returns dataframe indicating cluster number per document\n",
        "\n",
        "    km = KMeans(n_clusters=k, random_state =89)\n",
        "    km.fit(tfidf_matrix)\n",
        "    clusters = km.labels_.tolist()\n",
        "\n",
        "    Dictionary={'Doc Name':titles, 'Cluster':clusters,  'Text': final_processed_text}\n",
        "    frame=pd.DataFrame(Dictionary, columns=['Cluster', 'Doc Name','Text'])\n",
        "\n",
        "    print(\"Top terms per cluster:\")\n",
        "    #sort cluster centers by proximity to centroid\n",
        "    order_centroids = km.cluster_centers_.argsort()[:, ::-1] \n",
        "\n",
        "    terms_dict=[]\n",
        "\n",
        "    #dictionary to store clusters and respective titles\n",
        "    cluster_terms={}\n",
        "    cluster_title={}\n",
        "\n",
        "    #note doc2vec clusters will not have individual words due to the vector representation\n",
        "    #is based on the entire document not indvidual words. As a result, there won't be individual\n",
        "    #word outputs from each cluster.  \n",
        "\n",
        "    for i in range(k):\n",
        "        print(\"Cluster %d:\" % i),\n",
        "        temp_terms=[]\n",
        "        temp_titles=[]\n",
        "        temp=frame[frame['Cluster']==i]\n",
        "        docs_list = temp['Doc Name'].tolist()\n",
        "        cluster_matrix = tfidf_matrix.loc[docs_list]\n",
        "\n",
        "        df1 = pd.DataFrame(columns = ['term', 'Average tf-idf score'])\n",
        "\n",
        "        for ind in order_centroids[i, :20]:\n",
        "            #print(' %s' % terms[ind])\n",
        "            #print(round(cluster_matrix.iloc[:,ind].mean(),5))\n",
        "            terms_dict.append(terms[ind])\n",
        "            temp_terms.append(terms[ind])\n",
        "                        \n",
        "            df1 = df1.append({'term': terms[ind], 'Average tf-idf score':round(cluster_matrix.iloc[:,ind].mean(),4) }, ignore_index = True)\n",
        "        cluster_terms[i]=temp_terms\n",
        "        print(df1)\n",
        "        print(\"\\n\")\n",
        "\n",
        "        #print(\"Cluster %d titles:\" % i, end = '')\n",
        "        temp=frame[frame['Cluster']==i]\n",
        "        print('Cluster size: ', len(temp['Doc Name']))\n",
        "        # for title in temp['Doc Name']:\n",
        "        #     print(' %s' % title, end='\\n')\n",
        "        #     temp_titles.append(title)\n",
        "        # cluster_title[i]=temp_titles\n",
        "        print(\"\\n\")\n",
        "\n",
        "\n",
        "    return cluster_title,clusters,frame"
      ],
      "metadata": {
        "id": "Zx0qotJiS0CF"
      },
      "execution_count": 74,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "titles = list(range(10000))"
      ],
      "metadata": {
        "id": "gMkJllzAS0GK"
      },
      "execution_count": 34,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "tfidf_matrix, terms = tfidf(final_processed_text, titles, ngram_range = (1,1))"
      ],
      "metadata": {
        "id": "tfYdCq3VS0Kt"
      },
      "execution_count": 69,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "k_list = [5, 8, 10]\n",
        "\n",
        "#silhoutte score list\n",
        "sil_scores = []\n",
        "\n",
        "for k in k_list:\n",
        "    km = KMeans(n_clusters=k, random_state =89)\n",
        "    km.fit(tfidf_matrix)\n",
        "    labels = km.labels_.tolist()\n",
        "    score = silhouette_score(tfidf_matrix, labels)\n",
        "    sil_scores.append(score)"
      ],
      "metadata": {
        "id": "tyxLNJvre6uK"
      },
      "execution_count": 37,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "k_list2 = [4, 6, 7, 9]\n",
        "\n",
        "#silhoutte score list\n",
        "#sil_scores2 = []\n",
        "\n",
        "for k in k_list2:\n",
        "    km = KMeans(n_clusters=k, random_state =89)\n",
        "    km.fit(tfidf_matrix)\n",
        "    labels = km.labels_.tolist()\n",
        "    score = silhouette_score(tfidf_matrix, labels)\n",
        "    sil_scores.append(score)\n",
        "    k_list.append(k)"
      ],
      "metadata": {
        "id": "u0rS6qzHiqa8"
      },
      "execution_count": 41,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "k_list3 = [2, 3, 11, 12]\n",
        "\n",
        "#silhoutte score list\n",
        "#sil_scores2 = []\n",
        "\n",
        "for k in k_list3:\n",
        "    km = KMeans(n_clusters=k, random_state =89)\n",
        "    km.fit(tfidf_matrix)\n",
        "    labels = km.labels_.tolist()\n",
        "    score = silhouette_score(tfidf_matrix, labels)\n",
        "    sil_scores.append(score)\n",
        "    k_list.append(k)"
      ],
      "metadata": {
        "id": "Rd4PTUumpjcg"
      },
      "execution_count": 61,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "sil = list(zip(k_list,sil_scores))\n",
        "sil_df = pd.DataFrame(sil, columns = ['k_list', 'sil_scores'])\n",
        "sil_df = sil_df.sort_values(by=['k_list'])"
      ],
      "metadata": {
        "id": "KVTL_1nXoe5f"
      },
      "execution_count": 63,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "# plot silhouette score\n",
        "plt.figure(figsize=(12, 4))\n",
        "plt.scatter(sil_df['k_list'], sil_df['sil_scores'])\n",
        "plt.plot(sil_df['k_list'], sil_df['sil_scores'])\n",
        "plt.xlabel(\"clusters\")\n",
        "plt.ylabel(\"sillhouette score\")\n",
        "plt.show()"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 279
        },
        "id": "XAH1xoHdhUno",
        "outputId": "9db3dbda-1327-4940-b518-bda37e598cb3"
      },
      "execution_count": 65,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "<Figure size 864x288 with 1 Axes>"
            ],
            "image/png": "iVBORw0KGgoAAAANSUhEUgAAAuMAAAEGCAYAAADPHJsIAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4yLjIsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+WH4yJAAAgAElEQVR4nOzdd3xV9f3H8deHLPYIAWQTIExlSAqoiFtQq6jFFm3dLbVqly0Knf7scGCt1tFq1Wqto4qoOFHBvSCIskPCTpgBwszO5/fHPdAQAlwgNzfJfT8fj/vIOd/zPed+jhfDh+/9nO/X3B0REREREal5DaIdgIiIiIhIrFIyLiIiIiISJUrGRURERESiRMm4iIiIiEiUKBkXEREREYmS+GgHEE0pKSnerVu3aIchIiIiIvXcnDlz8ty9TeX2mE7Gu3XrRkZGRrTDEBEREZF6zsxWVdUe0TIVMxttZplmlm1mE6s4nmRm/w2Of2Fm3SocmxS0Z5rZqKCtt5l9VeG13cx+FhxLNrN3zCwr+NkqkvcmIiIiInK0IpaMm1kc8CBwDtAPuNTM+lXqdi2w1d17An8F7gzO7QeMA/oDo4GHzCzO3TPdfZC7DwKGALuBl4JrTQRmuHsaMCPYFxERERGptSI5Mj4UyHb35e5eDDwHjKnUZwzwZLA9BTjDzCxof87di9x9BZAdXK+iM4Bl7r6qims9CVxYrXcjIiIiIlLNIpmMdwTWVNjPCdqq7OPupcA2oHWY544Dnq2w387d1wXb64F2VQVlZuPNLMPMMjZt2hT+3YiIiIiIVLM6ObWhmSUCFwAvVHXc3R3wAxx7xN3T3T29TZv9HmgVEREREakxkZxNJRfoXGG/U9BWVZ8cM4sHWgCbwzj3HOBLd99QoW2DmbV393Vm1h7YWD23ISIiEhtenpvL5OmZrM0voEPLRkwY1ZsLB1f+YlpEqlMkR8ZnA2lmlhqMZI8DplXqMw24MtgeC8wMRrWnAeOC2VZSgTRgVoXzLmXfEpXK17oSeKXa7kRERKSee3luLpOmzic3vwAHcvMLmDR1Pi/PrTyOJiLVKWLJeFADfiMwHVgMPO/uC83sNjO7IOj2GNDazLKBmwhmQHH3hcDzwCLgLeAGdy8DMLMmwFnA1EpveQdwlpllAWcG+yIiIhKGydMzKSgp26etoKSMydMzoxSRSGyI6KI/7v4G8Ealtt9V2C4ELjnAuX8C/lRF+y5CD3lWbt9MaIYVEREROUy5+QVVtq89QLuIVI86+QCniIiIVI+i0jIe/Wg5ZlUfT0poQN7OopoNSiSGKBkXERGJQeXlzstzcznjLx/wx9cX06tdM5Li900L4hsYJWXOWfd8wKtfryX0WJeIVCcl4yIiIjHmo6xNfPP+j/nZf7+iRaME/nPtMKb/bCR3fmsAHVs2woCOLRtx9yUDeeunJ9OldRN+/OxcrvvPHDbuKIx2+CL1isXyv3LT09M9IyMj2mGIiIjUiAW527jzrSV8lJVHp1ahqQvPH9CBBg0OUKMSKC0r59GPV3DPO0tpnBjHref3Z8ygDtiBaltEZD9mNsfd0/drVzKuZFxEROq3NVt2c/fbmbzy1VpaNU7gxtPT+N7wLiTFxx3WdbI37mTClK+ZuzqfM/u2488XHUvb5g0jFLVI/aJkvApKxkVEpD7buquYB97L5qnPVmEG145I5bpTe9C8YcIRX7Os3PnXJyuYPD2TpPgG/P78/lx8fEeNkoscgpLxKigZFxGR+qiwpIzHP1nB399fxq6iUi4Z0pmfn9WLY1pU3yj28k07uXnKPDJWbeX0Pm3580XHVev1RapTbVhdVsl4FZSMi4hIfVJW7rw4J4d73lnK+u2FnNm3LTeP7kOvds0i9n5PfLqSydOXkBDXgN9+sx+XDOmkUXKpVfasLltxUatGCXHcfvFxNZqQKxmvgpJxERGpD9ydmUs2cudbS1i6YSeDOrdk0jl9GNZ9vzXyImJl3i5unjKPWSu3cEqvNtx+8XF0aNmoRt5b5GDcneF/nsGGHfvPld+xZSM+mXh6jcVyoGQ8oitwioiISGTNXb2V299cwqwVW0hNacJD3z2ec449pkZHp7ulNOG58cN56vNV3PHmEkb99UN+fV5fvvONzhollxpVXFrOgrXbmLNyKxmrtjBn1VbydhZX2be2rC6rZFxERKQOWpG3i8nTl/DG/PWkNE3kD2P6M25oFxLiorOESIMGxpUnduO03m25+cWvmTh1Pq/PX8cdwdzlIpGQv7uYOau2krFqK3NWbuXrnHyKSssB6JLcmJFpbZi5ZCP5BSX7nVtbvr1RmYrKVEREpA7ZtKOIv83I4tlZq0mMb8APTu7OD0Z2p2lS7RlfKy93np61mtvfWIwBvzqvL5cN7aJRcjkq7s6qzbtDifeqLWSs3ErWxp1AaLXY/h1bkN61FeldWzGkWyvaNgs9UKya8VpMybiIiNQVu4pK+edHy/nnh8spLC3n0qGd+ckZaXsTjtpozZbd3PLiPD5dtpmTerbmjosH0Dm5cbTDkjqiuLSchWu3kbG35CSfvJ2h2u/mDeMZ0rUV6d2SGdK1FQM7taRR4oHnzddsKrWUknEREantSsrKeW72Gu57N4u8nUWcc+wxTBjVm+5tmkY7tLC4O8/OWsOf31hMuTuTzunDd4d1PeSqnxJ7tu0uYc7qLUHyvZWv1+xbcrJnxDu9azJpbZvWuT9DSsaroGRcRERqK3fnrQXrmTw9k+V5uxjaLZmJ5/bh+C6toh3aEcnNL2Dii/P4KCuP4d2TuetbA+nSWqPkseqwSk66tqoXK70qGa+CknEREamNZq3Ywu1vLmbu6nzS2jblltF9OKNv2zpfc+3uPJ+xhj++tpjScueW0b254oRudW6EUw7fnpKTOau27h35rlhycnyQeKd3Sz5kyUldpWS8CkrGRUSkNsnasIM731rCu4s30q55Ejed1YtvHd+J+CjNkBIpa/MLmDR1Ph8s3cTQbsncNXYA3VKaRDssqUb1veTkSCgZr4KScRERqQ3Wbyvk3neX8nzGGpokxnPdqT245qTUejk6uIe7M2VODre9toiSsnImjOrDVSd2Iy4GkrL6xt1ZvWX33sQ7Y+WWfUtOOjQnvVtyvSo5ORJKxqugZFxERKJpe2EJ/3h/GY9/soKycufy4d248fSeJDdJjHZoNWb9tkJ+9dJ8Zi7ZSHrXVtw1dkCdeTg1Vh2s5KTZnllOurZiSNdkBnWunyUnR0LJeBWUjIuISDQUlZbxn89X88DMLLbuLuGCgR345dm9Y/aBRnfnpbm53DptIUWl5fzy7N5cMyJVo+S1xLbdJXy5OjS94OyVW5mXk09hSajkpHNyI77RNTnmSk6OhJLxKigZFxGRmlRe7rw6by13v53Jmi0FnNSzNRNH9+W4Ti2iHVqtsHF7Ib96aQHvLt7A4C4tmTx2ID3bapS8uh1szu3KJSdzVm1h6YZ9S06GdE0mvVto9DtWS06OhJLxKigZFxGRmvJJdh63v7mYBbnb6du+ORPP6cPItJQ6P0NKdXN3pn29lt9PW8ju4jJuOqsX3x+RWu8eYo2WqlajTIprwDnHHUNhSblKTiJIyXgVlIyLiEikLVq7nTveWsKHSzfRsWUjfnF2Ly4c1FFf5R/Cxh2F/PblBUxfuIGBnVow+ZKB9GrXLNph1Xkn3j6DtdsKqzzWObkR6V2Tg5UtW9GrbTP9Oa1GSsaroGRcREQiJWfrbu55eykvfZVL84YJ3HhaTy4/oSsNEzSyGC5357V56/j9tIXsLCzlp2em8cOR3TVKfhj2LK7zUdYmPsrK4+1FGw7Yd+Ud59VgZLHnQMl4fDSCERERqa/ydxfz4HvZPPnpKjAYP7I715/SkxaNE6IdWp1jZpw/sAMn9GjN719ZyOTpmaFVSS8ZQJ9jmkc7vForf3cxny7bzEdZeXyUtYmcrQUAdGrViMaJcewuLtvvnI4tG9V0mBJQMi4iIlINCkvKeOLTlTz0XjY7ikr51vGduOmsXnRQknPUUpom8eB3j+fceev43SsLOP/+j/nx6Wn86NQeJGiUnOLScr5cvZWPs/L4KDuP+Tn5lDs0S4rnhB6t+eHI7pyc1oaurRvzyldr96sZb5QQx4RRvaN4B7EtomUqZjYauA+IAx519zsqHU8C/g0MATYD33H3lcGxScC1QBnwE3efHrS3BB4FjgUcuMbdPzOzW4EfAJuCy//K3d84WHwqUxERkaNVVu5M/TKHe95ZyrpthZzWuw23nNNHI7cRsnlnEbe+uohXv15L/w7NmTx2IP06xNZ/a3dn2aadwch3Hp8v38zu4jLiGhiDO7dkRFoKJ6elMLBTyypLeg42m4pETo3XjJtZHLAUOAvIAWYDl7r7ogp9rgcGuPt1ZjYOuMjdv2Nm/YBngaFAB+BdoJe7l5nZk8BH7v6omSUCjd09P0jGd7r73eHGqGRcRESOlLvzfuYm7nxrCUvW72BgpxZMPKcvJ/RoHe3QYsJbC9bzm5cXkL+7mBtP78n1p/YkMb7+jpJv3lnEx9l5fJyVx8fZeawLHsJMTWnCiJ6h5Ht4j9Y0b6hyqNoqGjXjQ4Fsd18eBPAcMAZYVKHPGODWYHsK8ICF5ngaAzzn7kXACjPLBoaa2SJgJHAVgLsXA8URvAcREZH9fL0mn9vfXMzny7fQtXVjHrhsMOcd117TFNag0ccew7DUZP7v1YXc+24W0xduYPLYARzbsX7M2V5YUsacVVv5MGsTH2flsXDtdgBaNEpgRM8URqSlMKJnCp2TY3OhqPokksl4R2BNhf0cYNiB+rh7qZltA1oH7Z9XOrcjUECoDOVfZjYQmAP81N13Bf1uNLMrgAzgF+6+tXJQZjYeGA/QpUuXo7pBERGp3yp/nX/1Sd2Yuyaf1+etI7lJIree34/LhnWt1yOytVmrJoncO24w5x7Xnl+/vIALH/yE60/twY2np9W5z8TdWbJ+x96671krNlNYUk5CnHF8l1ZMGNWbET1TOLZjC61MWs/UtQc444HjgR+7+xdmdh8wEfgt8HfgD4TqyP8A/AW4pvIF3P0R4BEIlanUUNwiIlLHVF4cJTe/gD++vpjEuAb8+PSejB/ZnWYqCagVzu5/DENTk7nttUX8bWZ2aJT8kgEM6NQy2qEd1MbthXwUlJ18nJ3Hph2hxXbS2jbl0qFdODkthWGprWmSVNfSNTkckfx0c4HOFfY7BW1V9ckxs3igBaEHOQ90bg6Q4+5fBO1TCCXjuPveiTPN7J/Aa9V2JyIiEnMmT8/cZ8aJPZKbJPKLszXzRG3TsnEi93x7EN8c0J5JU+dz0UOfct0p3fnJGWkkxdeOud0Lisv4YsXm0Oh3Vh6ZG3YA0LpJ4t6ykxFpKbRvoRl4Ykkkk/HZQJqZpRJKpMcBl1XqMw24EvgMGAvMdHc3s2nAM2Z2D6EHONOAWcEDnGvMrLe7ZwJnENSgm1l7d18XXPciYEEE701EROq5tfkFVbZv2F716oVSO5zepx1v/zyZP762iAffW8bbCzcw+ZKBDOpc86Pk5eXOonXb99Z9Z6zcSnFZOYnxDRjaLZmLju/IyWkp9D2muVa6jGERS8aDGvAbgemEpjZ83N0XmtltQIa7TwMeA54KHtDcQihhJ+j3PKFEuxS4wd33DE/8GHg6mEllOXB10H6XmQ0iVKayEvhhpO5NRETqv7bNk9iwvWi/ds0bXvu1aJTA5EsGcu6A9vxq6nwufugTfjCyOz8/s1fEV0Bdm1/Ax1l5fJi1iU+XbWbLrtA8E32OacZVJ3VjRM8UhqYmayVW2Sui84zXdpraUEREquLunPu3j1i8bsc+7Y0S4rj94uM0J3Mdsr2whD+/vpjnZq+hR5smTL5kIMd3aVVt199ZVMrnyzbzcXYoAV++KTSnRNtmSYxIS2FkWhtO6plCm2ZJ1faeUjdFY2pDERGROumlubksXreDiwZ3ZNaKLVocpQ5r3jCBO741gHOPa8/EF+cx9u+fcu2IVH5xdu8jGp0uK3fm5eSHHrzMyuPL1VspLXcaJjRgePfWXDa0CyentaFXu6aa6lLCopFxjYyLiEgFG3cUctY9H9KzbVNe+OEJquWtR3YUlnD7m0t45ovVdE9pwl1jB5DeLfmQ563evJuPskN1359k57G9sBQzOLZDi72rXQ7p2qrWPCgqtZNGxkVERMLw+1cWUlBSxp3fGqBEvJ5p1jCBP190HOcd156bp8zjkoc/Y2RaG5Zu2MH6bYV7v/04rU9bPluWt3fawVWbdwPQoUVDzjm2PSPSUjipZwrJTRKjfEdSH2hkXCPjIiISeGP+Oq5/+ktuHt2b60/tGe1wJIJ2FpVy3VNz+Dg7b5/2PZUl7tAkMY4TerTm5LQ2jEhLoXtKE5WeyBHTyLiIiMhBbN1VzO9eWcBxHVsw/uTu0Q5HIqxpUjwr8nbt1+4OzZLieeyqbzC4S0sS4urWSp5S9ygZFxERAW57bRH5u0v49zXDiFcCFhMONJf8zqJShqYeupZcpDrot42IiMS8mUs28NLcXK4/tQf9OjSPdjhSQw40Z7zmkpeapGRcRERi2vbCEn41dQG92jXlhtNVJx5LJozqTaNK0xs2SohjwqjeUYpIYpHKVEREJKbd/sYSNu4o5B+Xn6Sp6WLMnjnjJ0/P1FzyEjVKxkVEJGZ9mp3Hs7NWM35kdwZ1bhntcCQKLhzcUcm3RJXKVEREJCbtLi7llqnz6Na6MT8/s1e0wxGRGKWRcRERiUl3T1/Kmi0F/Hf8cBolqjxFRKJDI+MiIhJz5qzawr8+XcHlw7syrHvraIcjIjFMybiIiMSUwpIybp4yjw4tGnHLOX2iHY6IxDiVqYiISEz524wslm3axZPXDKVpkv4aFJHo0si4iIjEjAW523j4w+WMHdKJU3q1iXY4IiJKxkVEJDaUlJUzYco8kpsk8tvz+kU7HBERQGUqIiISI/7x/jIWr9vOw5cPoUXjhGiHIyICaGRcRERiwNINO7h/ZjbfHNCeUf2PiXY4IiJ7KRkXEZF6razcuXnKPJokxfF/F/SPdjgiIvtQmYqIiNRr//pkBV+tyee+cYNo3TQp2uGIiOxDI+MiIlJvrczbxeTpmZzZty0XDOwQ7XBERPajZFxEROql8nLnlhfnkRjXgD9eeBxmFu2QRET2o2RcRETqpadnreaLFVv4zTf7ckyLhtEOR0SkSodMxs2sl5nNMLMFwf4AM/tN5EMTERE5Mrn5BdzxxmJG9Ezh2+mdox2OiMgBhTMy/k9gElAC4O7zgHGRDEpERORIuTu/mjofB26/WOUpIlK7hZOMN3b3WZXaSiMRjIiIyNF68ctcPli6iZtH9aZzcuNohyMiclDhJON5ZtYDcAAzGwusC+fiZjbazDLNLNvMJlZxPMnM/hsc/8LMulU4NilozzSzURXaW5rZFDNbYmaLzeyEoD3ZzN4xs6zgZ6twYhQRkfpj4/ZCbnt1IeldW3HFCd2iHY6IyCGFk4zfADwM9DGzXOBnwHWHOsnM4oAHgXOAfsClZtavUrdrga3u3hP4K3BncG4/QqUw/YHRwEPB9QDuA95y9z7AQGBx0D4RmOHuacCMYF9ERGKEu/PbVxZQWFrOnWMH0KCBylNEpPY7aDIeJMDXu/uZQBugj7uPcPdVYVx7KJDt7svdvRh4DhhTqc8Y4MlgewpwhoWK+8YAz7l7kbuvALKBoWbWAhgJPAbg7sXunl/FtZ4ELgwjRhERqSfemL+e6Qs3cNNZvejRpmm0wxERCctBk3F3LwNGBNu73H3HYVy7I7Cmwn5O0FZlH3cvBbYBrQ9ybiqwCfiXmc01s0fNrEnQp5277ymfWQ+0qyooMxtvZhlmlrFp06bDuB0REamttuwq5nevLOC4ji34/ojUaIcjIhK2cMpU5prZNDO73Mwu3vOKeGRViweOB/7u7oOBXVRRjuLuTlDjXsWxR9w93d3T27RpE9FgRUSkZtz26kK2FZRw19gBxMdpCQ0RqTviw+jTENgMnF6hzYGphzgvF6g4uWunoK2qPjlmFg+0CN7rQOfmADnu/kXQPoX/JeMbzKy9u68zs/bAxjDuTURE6rgZizfw8ldr+ekZafRt3zza4YiIHJZDJuPufvURXns2kGZmqYQS6XHAZZX6TAOuBD4DxgIz3d3NbBrwjJndA3QA0oBZ7l5mZmvMrLe7ZwJnAIsqXeuO4OcrRxi3iIjUEdsLS/j1Swvo3a4ZN5zWM9rhiIgctkMm42bWCbgfOClo+gj4qbvnHOw8dy81sxuB6UAc8Li7LzSz24AMd59G6EHMp8wsG9hCsJhQ0O95Qol2KXBDUL8O8GPgaTNLBJYDe/6xcAfwvJldC6wCvh3WfwEREamzbn9jMRt3FPLw5UNIjFd5iojUPRYqrz5IB7N3gGeAp4Km7wHfdfezIhxbxKWnp3tGRka0wxARkSPwSXYe3330C344sjuTzu0b7XBERA7KzOa4e3rl9nCGEdq4+7/cvTR4PUFomkMREZGo2FVUysSp80hNacLPz+oV7XBERI5YOMn4ZjP7npnFBa/vEXrIUkREJComT89kzZYC7vzWABomxB36BBGRWiqcZPwaQvXX64F1hB60PNKHOkVERI5KxsotPPnZSq48oStDU5OjHY6IyFEJZzaVVcAFNRCLiIjIQRWWlHHzi/Po0KIRN4/uE+1wRESO2iFHxs3sSTNrWWG/lZk9HtmwRERE9nffjCyWb9rF7RcfR5OkcJbKEBGp3cIpUxng7vl7dtx9KzA4ciGJiIjsb37ONh75cDnfTu/EyF6aR0BE6odwkvEGZtZqz46ZJRPeyp0iIiLVori0nAlTvqZ1k0R+fV6/aIcjIlJtwkmq/wJ8ZmYvAEboAc4/RTQqERGRCv7xwTKWrN/BP69Ip0WjhGiHIyJSbcJ5gPPfZpYBnB40Xezuiw52joiISHXJXL+D+2dmcf7ADpzVr120wxERqVaHTMbNrAewzN0XmdmpwJlmtrZiHbmIiEgklJaVc/OUr2nWMIFbz1d5iojUP+HUjL8IlJlZT+BhoDPwTESjEhERAR7/ZAVf52zj1gv607ppUrTDERGpduEk4+XuXgpcDDzg7hOA9pENS0REYt2KvF385e2lnNm3HecP0F87IlI/hZOMl5jZpcAVwGtBm56eERGRiCkvd255cR6J8Q3400XHYmbRDklEJCLCScavBk4A/uTuK8wsFXgqsmGJiEgse/qLVcxasYXfntePds0bRjscEZGICWc2lUXATyrsrwDujGRQIiISu3K27uaON5dwcloKl6R3inY4IiIRFc7IuIiISI1wdyZNnY8Df77oOJWniEi9p2RcRERqjSlzcvgoK4+J5/Shc3LjaIcjIhJxYSfjZqbfiiIiEjEbtxfyh9cWMbRbMt8b1jXa4YiI1IhDJuNmdqKZLQKWBPsDzeyhiEcmIiIxw935zcsLKCot545vHUeDBipPEZHYEM7I+F+BUcBmAHf/GhgZyaBERCS2vD5/HW8v2sBNZ/Wie5um0Q5HRKTGhFWm4u5rKjWVRSAWERGJQVt2FfP7VxYyoFMLrh2RGu1wRERq1CGnNgTWmNmJgJtZAvBTYHFkwxIRkVjxf68uZHthCU+PHUZ8nOYVEJHYEs5vveuAG4COQC4wCLg+kkGJiEhseHfRBl75ai03nNaTPsc0j3Y4IiI1LpyR8d7u/t2KDWZ2EvBJZEISEZFYsK2ghF+/PJ8+xzTj+lN7RjscEZGoCGdk/P4w20RERML259cXs2lHEXeNHUBivMpTRCQ2HXBk3MxOAE4E2pjZTRUONQfiIh2YiIjUXx9lbeK/GWu47pQeDOjUMtrhiIhEzcGGIhKBpoQS9mYVXtuBseFc3MxGm1mmmWWb2cQqjieZ2X+D41+YWbcKxyYF7ZlmNqpC+0ozm29mX5lZRoX2W80sN2j/yszODSdGERGpWbuKSpn44ny6pzThZ2emRTscEZGoOuDIuLt/AHxgZgXuflfFY2Z2CZB1sAubWRzwIHAWkAPMNrNp7r6oQrdrga3u3tPMxgF3At8xs37AOKA/0AF418x6ufueKRVPc/e8Kt72r+5+98HiEhGR6Jo8PZO12wp44Ycn0DBBX7SKSGwLp0hvXBVtk8I4byiQ7e7L3b0YeA4YU6nPGODJYHsKcIaZWdD+nLsXufsKIDu4noiI1GGzV27hyc9WcuUJ3UjvlhztcEREou5gNePnAOcCHc3sbxUONQdKw7h2R6DiYkE5wLAD9XH3UjPbBrQO2j+vdG7HYNuBt83MgYfd/ZEK/W40syuADOAX7r61ivsaD4wH6NKlSxi3ISIi1aGwpIxbpsyjY8tGTBjVO9rhiIjUCgcbGV9LKKktBOZUeE0DRh3kvEgb4e7HA+cAN5jZyKD970APQvOgrwP+UtXJ7v6Iu6e7e3qbNm1qJGAREYF7381ied4u7rh4AE2SwplZV0Sk/jtYzfjXwNdm9kzQr4u7Zx7GtXOBzhX2OwVtVfXJMbN4oAWw+WDnuvuenxvN7CVC5SsfuvuGPZ3N7J/Aa4cRq4iIRNC8nHwe+XAZ30nvzIi0lGiHIyJSa4RTMz4a+Ap4C8DMBpnZtDDOmw2kmVmqmSUSqj2vfN404Mpgeyww0909aB8XzLaSCqQBs8ysiZk1C+JoApwNLAj221e47kV72kVEJLqKS8u5eco82jRL4lfn9Y12OCIitUo43xPeSmj0+X0Ad/8qSJAPKqgBvxGYTmhe8sfdfaGZ3QZkuPs04DHgKTPLBrYQPCwa9HseWESoPv0Gdy8zs3bAS6FnPIkHnnH3t4K3vMvMBhGqKV8J/DCMexMRkQh76P1slqzfwaNXpNOiUUK0wxERqVXCScZL3H1bkADv4eFc3N3fAN6o1Pa7CtuFwCUHOPdPwJ8qtS0HBh6g/+XhxCQiIjUnc/0OHnwvmzGDOnBmv3bRDkdEpNYJJxlfaGaXAXFmlgb8BPg0smGJiEhdV1pWzs1TvqZ5wwR+f37/aIcjIlIrhVMz/mNCi+8UAc8SWoHzZ5EMSkRE6r7HPl7B1znb+L8x/UlukhjtcEREaqVDjoy7+27g18FLRETkkJZv2sk975JF7isAACAASURBVCzl7H7tOO+49oc+QUQkRh0yGTez96iiRtzdT49IRCIiUqeVlzu3vDiPpPgG/PHCY6n0zJGIiFQQTs34LytsNwS+RXgrcIqISAz6zxermL1yK5PHDqBt84bRDkdEpFYLp0xlTqWmT8xsVoTiERGROmzNlt3c8eYSRvZqw9ghnaIdjohIrRdOmUpyhd0GwBBCK2WKiIjs5e786qX5GPDni1SeIiISjnDKVOYQqhk3QuUpK4BrIxmUiIjUPS9k5PBRVh5/GNOfTq0aRzscEZE6IZwylUOutikiIrFtw/ZC/vD6IoamJvPdYV2jHY6ISJ0RTplKAvAjYGTQ9D7wsLuXRDAuERGpI9ydX7+0gOLScu781gAaNFB5iohIuMJZ9OfvhOrEHwpeQ4I2ERERXp23jncXb+CXZ/cmNaVJtMMREalTwqkZ/4a7D6ywP9PMvo5UQCIiUnds3lnErdMWMrBzS64ZoapGEZHDFc7IeJmZ9dizY2bdgbLIhSQiInXFra8uYkdhCZPHDiBO5SkiIoctnJHxCcB7Zrac0IwqXYGrIxqViIjUem8vXM+rX6/lprN60atds2iHIyJSJ4Uzm8oMM0sDegdNme5eFNmwRESkNttWUMJvXl5An2Oa8aNTexz6BBERqVI4I+MQemizW9B/kJnh7v+OWFQiIlKr/en1RWzeVcxjV36DhLhwKh5FRKQq4Uxt+BTQA/iK/9WKO6BkXEQkhrw8N5fJ0zPJzS8A4Iy+bTmukxZkFhE5GuGMjKcD/dzdIx2MiIjUTi/PzWXS1PkUlPzv+f1PsvJ4eW4uFw7uGMXIRETqtnC+W1wAHBPpQEREpPZxd7buKubPbyzeJxEHKCwtZ/L0zChFJiJSPxxwZNzMXiVUjtIMWGRms4C9D266+wWRD09E6oI95Qtr8wvo0LIRE0b11mhpLVdQXEbeziI27ihk046i/712Fu23X1J24C9G1wYlKyIicmQOVqZyd41FISJ1VuXyhdz8AiZNnQ+ghLyGlZU7m3ftm0xvrJRk5wX7O4pK9zvfDFo3SaJNs9CrZ9tme7cfmJnF1t0l+53ToWWjmrg1EZF664DJuLt/UJOBiEjdNHl65n7lCwUlZdzy4jzeWbSBpknxNG0YH/p5sO1gXzNz7Mvd2VFUuu9odZBcb9y+70j2ll1FlFcxiN0sKZ42zZJIaZZE3w7NGdn0fwl322b/205unEj8Af77t26SuF/NeKOEOCaM6l1lfxERCc/BylR2ECpT2e8Q4O7ePGJRiUidkXuAMoWi0nIWr9/OzsJSdhWVsqs4vIV7k+IbHDBRP+B+FdtNEuNpEMEVIY+2NKeotIy8ncVVJNmF+45o7yiiqLR8v/MT4ow2QVLdsWVDBnVusXc/9GpI22ZJpDRNolFi3FHf7557UzmSiEj1OtjIuJZTE5FDapoUz84qSh46tmzEzF+cune/rNzZVRxKzHcWlrIj+LmraN/tncF+xX7rthWya9P/9ourSE6r0iQx7gBJewJNk/Ycq7y9f4LfMKEBZv9L7A9UmlPuzqm92+5fh125FntnEflVlHwAJDdJ3JtUd+vWJJRYN91/FLtFo4R9YqoJFw7uqORbRKSaHWxkPPlgJ7r7luoPR0Tqksc+XsHOolLizCirMPtpVeULcQ2M5g0TaN4wAY5yauri0vK9ifveV+H+Sfye7Yr98nbs3me/rKq6jkriGtg+o/Ir8nZRXLbvPwgKSsq46fmvqzy/UUIcbZuHkuqebZtyQo/WlUaxk2jbrCGtmyaqTEdEJMYc7AHOOYTKVKoaenGge0QiEpE64eW5ufzhtUWM7n8MZ/drx1/eWVpj5QuJ8Q1IjE+kVZPEo7qOu1NYUr5PQv+/RL2EnUVlQVtJ8LOMnUUlZG7YccBr3np+P9o0a7hPTXaTpHAXOxYRkVhzsDKV1JoMRETqjvczN/LLF75mePdk7h03iIYJcVw8pFO0wzpsZkajxDgaJcbRpllS2OeddMfMKmvlO7ZsxFUn6VeniIiE74Dfh5pZn+Dn8VW9wrm4mY02s0wzyzaziVUcTzKz/wbHvzCzbhWOTQraM81sVIX2lWY238y+MrOMCu3JZvaOmWUFP1uF959ARA7Hl6u38qP/fEmvds145Ip0GiYc/cOBdc2EUb1pVOm+NbOIiIgciYN9d3oTMB74SxXHHDj9YBc2szjgQeAsIAeYbWbT3H1RhW7XAlvdvaeZjQPuBL5jZv2AcUB/oAPwrpn1cvc90zGc5u55ld5yIjDD3e8IEv+JwC0Hi1FEDk/Whh1c88Rs2jZP4slrhobqv2OQZhYREZHqcrAylfHBz9OO8NpDgWx3Xw5gZs8BY4CKyfgY4NZgewrwgIWmBxgDPOfuRcAKM8sOrvfZQd5vDHBqsP0k8D5KxkWqTW5+AVc8PouEuAY8dc2wwyrrqI80s4iIiFSHQz62b2aXmFmzYPs3ZjbVzAaHce2OwJoK+zlBW5V93L0U2Aa0PsS5DrxtZnPMbHyFPu3cfV2wvR5od4D7GW9mGWaWsWnTpjBuQ0S27Crmise+YGdhKU9ePZQurRtHOyQREZF6IZw5tH7r7jvMbARwJvAY8I/IhnVQI9z9eOAc4AYzG1m5g7s7VS9YhLs/4u7p7p7epk2bCIcqUvftKirl6idms2ZrAY9emU6/DlrvS0REpLqEk4zvqdM+D3jE3V8HwplPLBfoXGG/U9BWZR8ziyc0+/Dmg53r7nt+bgReIlS+ArDBzNoH12oPbAwjRhE5iOLScn709JfMz8nngUsHM6x762iHJCIiUq+Ek4znmtnDwHeAN8wsKczzZgNpZpZqZomEHsicVqnPNODKYHssMDMY1Z4GjAtmW0kF0oBZZtakQslME+BsYEEV17oSeCWMGEXkAMrLnV++8DUfLt3E7Rcfx9n9j4l2SCIiIvVOOCtRfBsYDdzt7vnBqPOEQ53k7qVmdiMwHYgDHnf3hWZ2G5Dh7tMIlbw8FTyguYVQwk7Q73lCD3uWAje4e5mZtQNeCpaAjgeecfe3gre8A3jezK4FVgVxi8gRcHdue20R075ey82je/Odb3SJdkgiIiL1krkfeino+io9Pd0zMjIO3VEkxjwwM4u7317KtSNS+c15fQn+ASwiIiJHyMzmuHt65fZwyk1EJIY888Vq7n57KRcN7sivz1UiLiIiEklKxkVkr7cWrOM3L8/n1N5tuGvsABo0UCIuIiISSUrGRQSAT5fl8ZNnv2JQ55Y89N3jSYjTrwcREZFI09+2IsKC3G2M//ccurZuzONXfYPGieE82y0iIiJHS8m4SIxbmbeLq/41i+YN4/n3tUNp2TicZQRERESkOigZF4lhG7cXcvnjX1BW7vz72mG0b9Eo2iGJiIjEFH0XLRKjthWUcMXjs9i8s5hnfjCcnm2bRjskERGRmKORcZEYVFhSxg+ezGDZpp08fPkQBnVuGe2QREREYpJGxkViTGlZOT9+di6zV23hb+MGc3Jam2iHJCIiErM0Mi4SQ9ydX700n3cWbeDW8/tz/sAO0Q5JREQkpikZF4khd03P5PmMHH5yek+uPLFbtMMRERGJeUrGRWLEox8t5+/vL+OyYV34+Vm9oh2OiIiIoGRcJCZM/TKHP76+mHOOPYY/jDkWMy1zLyIiUhsoGRep595bspEJU+ZxYo/W3DtuEHENlIiLiIjUFkrGReqxOau28qOn59C3fTMevnwISfFx0Q5JREREKlAyLlJPLd2wg2uemM0xzRvyxNVDadYwIdohiYiISCVKxkXqoZytu7nisVkkxjfgqWuHkdI0KdohiYiISBWUjIvUM5t3FnHFY7PYVVzKv68ZSufkxtEOSURERA5AybhIPbKzqJSrn5hNbn4Bj135Dfq2bx7tkEREROQg4qMdgNRvL8/NZfL0TNbmF9ChZSMmjOrNhYM7Rjuseqm4tJzrnprDwrXbefh7QxiamhztkEREROQQlIxLxLw8N5dJU+dTUFIGQG5+AZOmzgdQQl7Nysudm57/io+z85g8dgBn9msX7ZBEREQkDCpTkYiZPD1zbyK+R0FJGZOnZ0YpovrJ3fm/Vxfy2rx1TDynD5ekd452SCIiIhImJeMSMbn5BYfVLkfm/pnZPPnZKn5wcio/HNk92uGIiIjIYVAyLtVuZ1Ep9767lAOt8xjXwHhu1mqKS8trNK766D+fr+Ked5Zy8fEdmXROXy1zLyIiUscoGZdqU1xazhOfrOCUu97j3nezOK5TC5Li9/0jlhjXgA4tGjJx6nxOu/t9nv5ilZLyI/TG/HX89pUFnN6nLXd+awANtMy9iIhInaMHOOWolZc7075ey1/eyWTNlgKGd0/m0dF9GNylVZWzqYwZ1IH3l27ivnez+PVLC3hwZjY/Oq0n307vpOXaw/Rpdh4/e+4rju/SigcvO56EOP27WkREpC4yd4/cxc1GA/cBccCj7n5HpeNJwL+BIcBm4DvuvjI4Ngm4FigDfuLu0yucFwdkALnu/s2g7QngFGBb0O0qd//qYPGlp6d7RkbGUd5l7HJ33l+6ibveymTxuu30a9+cm0f35pRebcIql3B3PszK4753l/Ll6nzat2jIj07twbfTO9MwQUn5gczP2ca4Rz6jU6vGPP/DE2jRWMvci4iI1HZmNsfd0yu3R2xkPEiYHwTOAnKA2WY2zd0XVeh2LbDV3Xua2TjgTuA7ZtYPGAf0BzoA75pZL3ffMzXHT4HFQOUVTSa4+5RI3ZP8z5ert3Lnm0v4YsUWuiQ35r5xgzh/QIfDKpUwM07p1YaRaSl8kr2Z+2Ys5XevLOTB97K57pQeXDq0i5LySlbk7eKqf82iZeNEnrxmqBJxERGROi6SZSpDgWx3Xw5gZs8BY4CKyfgY4NZgewrwgIWGVMcAz7l7EbDCzLKD631mZp2A84A/ATdFMH6pQvbGHUyensn0hRtIaZrIbWP6M+4bXUiMP/IyCTNjRFoKJ/VszWfLNnPvjCz+79VFPPT+Mq47pQeXDe1Co0Ql5Ru2F3L5Y1/gwFPXDuWYFg2jHZKIiIgcpUgm4x2BNRX2c4BhB+rj7qVmtg1oHbR/XuncPavE3AvcDDSr4j3/ZGa/A2YAE4Nkfh9mNh4YD9ClS5fDvKXYtW5bAfe+k8ULc9bQODGem87qxbUjUmmSVH1/hMyME3umcGLPFD5fvpn73s3iD68t4u/vL+OHI7vz3eFdaJwYm485bNtdwhWPzWLrrmKeHT+c7m2aRjskERERqQZ1KrMxs28CG919jpmdWunwJGA9kAg8AtwC3Fb5Gu7+SHCc9PT0yBXM1xP5u4v5+/vLeOLTlbjDVSemcsNpPWjdNCmi7zu8e2uGj2/NrBVbuG/GUv70xmL+8cEyxo/szveGd63WfwTUdgXFZVz75GyW5+3kX1cNZUCnltEOSURERKpJJDOaXKDiUoCdgraq+uSYWTzQgtCDnAc69wLgAjM7F2gINDez/7j799x9XdC3yMz+Bfyyum8olhQUl/GvT1fwj/eXsaOolIsGd+TnZ/aic3LjGo1jaGoyT39/OBkrt3DfjCxuf3MJD3+4nO+fnMoVJ3SjaT1PykvLyrnxmS+Zs3or9186mBFpKdEOSURERKpRxGZTCZLrpcAZhBLp2cBl7r6wQp8bgOPc/brgAc6L3f3bZtYfeIZQnXgHQmUnaRUe4CQYGf9lhdlU2rv7uqDm/K9AobtPPFiMmk1lf6Vl5TyfkcN9M5ayYXsRZ/Rpy4TRvelzTOVnZaNjzqqt/G1GFh8s3UTLxgn84OTuXHFCV5o1rH8PMro7E6bMY8qcHP5w4bFcPrxrtEMSERGRI1Tjs6kENeA3AtMJTW34uLsvNLPbgAx3nwY8BjwVPKC5hdAMKgT9nif0sGcpcEPFRPwAnjazNoABXwHXReTG6il3580F67l7eibL83YxpGsrHrjseL7RLTnaoe1jSNdWPHnNUL5ak8/fZmQxeXomj3y4nO+PSOXKk7rRvB4l5Xe8tYQpc3L42ZlpSsRFRETqqYjOM17baWQ85NPsPO58awlf52yjV7umTBjVhzP7tq0TS6vPywkl5e8u3kjzhvFcMyKVq09KpUWjup2UP/LhMv78xhIuH96V28b0rxOfhYiIiBzYgUbGlYzHcDK+IHcbd761hI+y8ujQoiE/P6sXFx/fibg6uKz6gtxt3Dcji3cWbaBZw3iuPimVa09KrZPzcE+Zk8MvX/ia845rz98uHVwnPw8RERHZl5LxKsRqMr4ybxd3v53Ja/PW0bJxAjee1pPvDe9aLxbYWbh2G/fPyOathetpmhTPVSd249oRqbRqkhjt0MIyY/EGxj81h+Hdk3n8qm+QFF/3PxMRERFRMl6lWEvGN+4o5P4Z2Tw7azUJcQ34/smp/GBk93pVZ73H4nXbuX9mFm/MX0+TxDiuPLEb3z+5O8m1OCnPWLmF7z76Bb3aNePZ8cPr/UwxIiIisUTJeBViJRnfXljCIx8s57GPV1BSVs6lQ7vw4zN60rZZ/V/BMXP9Du6fmcXr89fRKCGOy0/oyviTu0d8nvTDlbl+B5f841NSmibxwnUn1Lr4RERE5OgoGa9CfU/GC0vK+M/nq3jwvWy27i7h/IEd+MVZveiW0iTaodW4rA07uH9mNq/OW0vD+CApH9mdlFqQ9K7Zspux//gUgCnXnVjjc7mLiIhI5CkZr0J9TcbLyp2pX+Zw77tZ5OYXcHJaCreM7sOxHVtEO7Soy964kwffy+aVr3JJjG/A94Z1Zfwp3aP2LUHeziIu+cdnbN5ZxAvXnUjvY5pFJQ4RERGJLCXjVahvybi78+7ijUyevoSlG3YysFMLbhndhxN7atXGypZv2skD72XzyldriW9gfHdYV647pTttm9dcUr6zqJRLH/mcpRt28PT3h5Fey+Z0FxERkeqjZLwK9SkZn71yC3e+uYSMVVvpntKEX47qzTnHHqP5qQ9hZd4uHnwvm6lzc4lrYFw2tAvXndKDY1pENikvKi3jmidm8/nyLTxy+RDO6Nsuou8nIiIi0aVkvAr1IRlfsn47k9/KZMaSjbRtlsTPzuzFJemdSIhrEO3Q6pTVm3fz4HvZvPhlDg3MGDe0M9ed0oMOLRtV+3uVlTs/eXYur89fx18uGci3hnSq9vcQERGR2kXJeBXqcjKes3U397yzlJfm5tI0KZ4fndqDq09MpVGi5qU+Gmu27Oah97N5ISOUlF+S3onrT+tJx2pKyt2d372ykKc+X8Wvz+3LD0Z2r5brioiISO2mZLwKdTEZ37KrmAdmZvOfz1dhBled2I0fndqDlo1r7/zZdVHO1t38/f1lPJ+xBoCxQzpz/ak9jnqmk3vfXcq972bxw1O6M+mcvtURqoiIiNQBSsarUJeS8V1FpTz28Qoe+XA5u4tLuWRIZ352VhrtW1R/GYX8z9r8Av7+/jL+O3sN5e6MHdKJ60/tSZfWh5+UP/XZSn77ykLGDunE5LEDVM8vIiISQ5SMV6EuJOPFpeU8N3s1f5uRTd7OIkb1b8eEUb3p2VZT4NWkddsKePiD5TwzazVl5c7Fgzty4+k96do6vDnbX5u3lh8/O5fTe7fl4cuHEK+afhERkZiiZLwKtTkZLy93Xp23lr+8vZTVW3YzLDWZW87pw/FdWkU7tJi2YXsh//hgGc98sZrScufCQaGkPPUgCyl9nJXH1U/MYlDnlvz7mmGq6xcREYlBSsarUBuTcXfnw6w87nprCQvXbqdv++bcPLo3p/Zqo7KGWmTj9kIe+XA5//liFcWl5YwJkvIebZry8txcJk/PZG1+ASlNk9hWWEL3lCb8d/wJtGicEO3QRUREJAoOlIzHRyMYqdpXa/K5483FfL58C52TG3HfuEGcP6ADDRooCa9t2jZvyG++2Y8fntKDf360nKc+W8UrX+UyuHMrFqzdRlFpOQCbdhZhwKVDuygRFxERkf1oZLwWjIxnb9zJX97O5M0F60lpmsiPT0/j0qFdSIxXXXFdkbeziH9+tJxHPlhOVf9HdWzZiE8mnl7jcYmIiEjtoJHxWqBi+UKHlo34wcmpLFm/gxfm5NAwvgE/P7MX156cStMkfSx1TUrTJCad05eHP1he5fG1+QU1HJGIiIjUBcr6asjLc3OZNHU+BSVlAOTmF3Drq4uIa2BccUJXbjytJ62bJkU5SjlaHVs2IreKxDsSK3mKiIhI3ac6iBoyeXrm3kS8opQmifz+/P5KxOuJCaN60yhh39lSGiXEMWFU7yhFJCIiIrWZRsZryIHKFDbuKKrhSCSSLhzcEWCfcqQJo3rvbRcRERGpSMl4Demg8oWYceHgjkq+RUREJCwqU6khKl8QERERkco0Ml5DVL4gIiIiIpUpGa9BKl8QERERkYpUpiIiIiIiEiVKxkVEREREoiSiybiZjTazTDPLNrOJVRxPMrP/Bse/MLNuFY5NCtozzWxUpfPizGyumb1WoS01uEZ2cM3ESN6biIiIiMjRilgybmZxwIPAOUA/4FIz61ep27XAVnfvCfwVuDM4tx8wDugPjAYeCq63x0+BxZWudSfw1+BaW4Nri4iIiIjUWpEcGR8KZLv7cncvBp4DxlTqMwZ4MtieApxhZha0P+fuRe6+AsgOroeZdQLOAx7dc5HgnNODaxBc88KI3JWIiIiISDWJ5GwqHYE1FfZzgGEH6uPupWa2DWgdtH9e6dw905DcC9wMNKtwvDWQ7+6lVfTfh5mNB8YHuzvNLPMw7qm6pAB5UXhfqVn6nGODPufYoM+5/tNnHBui+Tl3raqxTk1taGbfBDa6+xwzO/VIruHujwCPVGtgh8nMMtw9PZoxSOTpc44N+pxjgz7n+k+fcWyojZ9zJMtUcoHOFfY7BW1V9jGzeKAFsPkg554EXGBmKwmVvZxuZv8JzmkZXONA7yUiIiIiUqtEMhmfDaQFs5wkEnogc1qlPtOAK4PtscBMd/egfVww20oqkAbMcvdJ7t7J3bsF15vp7t8LznkvuAbBNV+J4L2JiIiIiBy1iCXjQf32jcB0QjOfPO/uC83sNjO7IOj2GNDazLKBm4CJwbkLgeeBRcBbwA3uXnaIt7wFuCm4Vuvg2rVVVMtkpMboc44N+pxjgz7n+k+fcWyodZ+zhQaVRURERESkpmkFThERERGRKFEyLiIiIiISJUrGa5CZdTaz98xskZktNLOfRjsmiQwzizOzuWb2WrRjkcgws5ZmNsXMlpjZYjM7IdoxSfUzs58Hv68XmNmzZtYw2jHJ0TOzx81so5ktqNCWbGbvmFlW8LNVNGOUo3eAz3ly8Ht7npm9ZGYtoxkjKBmvaaXAL9y9HzAcuMHM+kU5JomMnxJ6cFnqr/uAt9y9DzAQfd71jpl1BH4CpLv7sUAcoZm8pO57AhhdqW0iMMPd04AZwb7UbU+w/+f8DnCsuw8AlgKTajqoypSM1yB3X+fuXwbbOwj95V3lSqFSd5lZJ+A84NFoxyKRYWYtgJEEsza5e7G750c3KomQeKBRsI5FY2BtlOORauDuHwJbKjWPAZ4Mtp8ELqzRoKTaVfU5u/vbFVZs/5zQ2jRRpWQ8SsysGzAY+CK6kUgE3AvcDJRHOxCJmFRgE/CvoBzpUTNrEu2gpHq5ey5wN7AaWAdsc/e3oxuVRFA7d18XbK8H2kUzGKkR1wBvRjsIJeNRYGZNgReBn7n79mjHI9XHzL4JbHT3OdGORSIqHjge+Lu7DwZ2oa+0652gZngMoX98dQCamNn3ohuV1IRgMUHN/VyPmdmvCZUPPx3tWJSM1zAzSyCUiD/t7lOjHY9Uu5OAC8xsJfAccLqZ/Se6IUkE5AA57r7nm60phJJzqV/OBFa4+yZ3LwGmAidGOSaJnA1m1h4g+LkxyvFIhJjZVcA3ge96LVhwR8l4DTIzI1Rjutjd74l2PFL93H2Su3dy926EHvSa6e4aSatn3H09sMbMegdNZxBaMVjql9XAcDNrHPz+PgM9qFufTQOuDLavBF6JYiwSIWY2mlAp6QXuvjva8YCS8Zp2EnA5odHSr4LXudEOSkSOyI+Bp81sHjCI/2/vbkKtLOI4jn9/FdFC4VLS5iKKdalQSlJctHAhKK1aVateoCCiRRAkvdAiF6mgm9SFi8IoKuhlUVARQdgirhQU6W0RgtygRQsx6AUXJf8Wz0iHi4LXjs29ne8HDgzzzDzMcOA8P+YZzsDuzuPRmLU3H+8B3wAnGJ6ZS+4obS1ekreBWeCWJD8leRTYC2xPcpLhrcjenmPUv3eR7/kQsBL4rOWww10HCWQJrM5LkiRJE8mVcUmSJKkTw7gkSZLUiWFckiRJ6sQwLkmSJHViGJckSZI6MYxL0oRK8mKSpy+j31SSJ67EmCRp0hjGJUmLNQUsKoxn4DNHkhbwh1GSJkSSh5IcT/JdkjcWXDuaZHMrr0oy38rrk3zVDsc4nmSG4TCUm1rdvtZuZ5KvW5tdrW5tkh+SvA7MAauTvJZkLsmJJE/9h9OXpCXpmt4DkCRdeUnWAy8Ad1XV6STXA09eQtfHgZer6s0k1wJXA88CG6pqY7v3DmAG2AIE+DDJVobj5GeAh6vqWJJNwHRVbWj9psY7S0laflwZl6TJsA14t6pOA1TVmUvsNws8n+QZYE1Vnb1Amx3t8y3D0fG3MoRwgB+r6lgrnwLWJTmY5G7g18ubiiT9fxjGJUkAf/HPM+G685VV9RZwD3AW+DjJtgv0DbCnqja2z81V9Wq79sfIvX4B7gCOMqy4vzL2WUjSMmMYl6TJ8DlwX5IbANo2lVHzwKZWvvd8ZZJ1wKmqOgB8ANwO/AasHOn7KfBIkhWtz3SSGxcOIMkq4Kqqep9hy8ydY5iXJC1r7hmXpAlQVd8neQn4Isk5hi0l8yNN9gPvJHkM+Gik/n7gwSR/Aj8Du6vqTJIvk8wBn1TVziS3AbNJAH4HHgDOLRjG579khgAAAE1JREFUNHBk5F9VnhvvLCVp+UlV9R6DJEmSNJHcpiJJkiR1YhiXJEmSOjGMS5IkSZ0YxiVJkqRODOOSJElSJ4ZxSZIkqRPDuCRJktTJ3wfs8KhcPkuLAAAAAElFTkSuQmCC\n"
          },
          "metadata": {
            "needs_background": "light"
          }
        }
      ]
    },
    {
      "cell_type": "code",
      "source": [
        "cluster_title,clusters,k_means_df = k_means(titles, tfidf_matrix, k =9)"
      ],
      "metadata": {
        "id": "uz21M_0BS0IW",
        "colab": {
          "base_uri": "https://localhost:8080/"
        },
        "outputId": "5a48a976-0d1c-4053-dd7d-f29e106b1ddb"
      },
      "execution_count": 75,
      "outputs": [
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "Top terms per cluster:\n",
            "Cluster 0:\n",
            "         term  Average tf-idf score\n",
            "0        taco                0.2656\n",
            "1       salsa                0.0804\n",
            "2     mexican                0.0684\n",
            "3        chip                0.0465\n",
            "4        fish                0.0462\n",
            "5        good                0.0452\n",
            "6        food                0.0447\n",
            "7   margarita                0.0438\n",
            "8       great                0.0315\n",
            "9   delicious                0.0306\n",
            "10    burrito                0.0291\n",
            "11     street                0.0280\n",
            "12      place                0.0280\n",
            "13       like                0.0273\n",
            "14   tortilla                0.0268\n",
            "15       back                0.0263\n",
            "16    amazing                0.0255\n",
            "17    chicken                0.0246\n",
            "18     really                0.0242\n",
            "19       time                0.0238\n",
            "\n",
            "\n",
            "Cluster size:  275\n",
            "\n",
            "\n",
            "Cluster 1:\n",
            "         term  Average tf-idf score\n",
            "0       pizza                0.3253\n",
            "1       crust                0.0574\n",
            "2        good                0.0446\n",
            "3       salad                0.0404\n",
            "4       place                0.0369\n",
            "5       great                0.0352\n",
            "6      cheese                0.0348\n",
            "7       style                0.0321\n",
            "8        like                0.0279\n",
            "9       louis                0.0274\n",
            "10       best                0.0259\n",
            "11    ordered                0.0250\n",
            "12      sauce                0.0245\n",
            "13       time                0.0245\n",
            "14    topping                0.0240\n",
            "15       thin                0.0235\n",
            "16  pepperoni                0.0234\n",
            "17    service                0.0221\n",
            "18      order                0.0219\n",
            "19       back                0.0216\n",
            "\n",
            "\n",
            "Cluster size:  569\n",
            "\n",
            "\n",
            "Cluster 2:\n",
            "        term  Average tf-idf score\n",
            "0       good                0.0540\n",
            "1    chicken                0.0538\n",
            "2      sauce                0.0387\n",
            "3    ordered                0.0367\n",
            "4      salad                0.0353\n",
            "5       like                0.0346\n",
            "6     cheese                0.0329\n",
            "7   sandwich                0.0309\n",
            "8     really                0.0306\n",
            "9       side                0.0292\n",
            "10      food                0.0284\n",
            "11      also                0.0283\n",
            "12     place                0.0279\n",
            "13     fried                0.0278\n",
            "14      dish                0.0253\n",
            "15     great                0.0249\n",
            "16      time                0.0240\n",
            "17    burger                0.0237\n",
            "18      back                0.0233\n",
            "19    flavor                0.0232\n",
            "\n",
            "\n",
            "Cluster size:  2243\n",
            "\n",
            "\n",
            "Cluster 3:\n",
            "         term  Average tf-idf score\n",
            "0       sushi                0.3096\n",
            "1        roll                0.1779\n",
            "2       place                0.0409\n",
            "3       great                0.0384\n",
            "4        good                0.0377\n",
            "5     service                0.0321\n",
            "6       fresh                0.0282\n",
            "7        fish                0.0270\n",
            "8        time                0.0255\n",
            "9      really                0.0247\n",
            "10  recommend                0.0247\n",
            "11      price                0.0240\n",
            "12       food                0.0236\n",
            "13       tuna                0.0234\n",
            "14       best                0.0223\n",
            "15    tempura                0.0218\n",
            "16       back                0.0213\n",
            "17       crab                0.0212\n",
            "18  delicious                0.0209\n",
            "19      spicy                0.0206\n",
            "\n",
            "\n",
            "Cluster size:  265\n",
            "\n",
            "\n",
            "Cluster 4:\n",
            "          term  Average tf-idf score\n",
            "0         beer                0.2132\n",
            "1    selection                0.0655\n",
            "2        great                0.0552\n",
            "3         good                0.0548\n",
            "4         food                0.0475\n",
            "5        place                0.0455\n",
            "6         nice                0.0304\n",
            "7         game                0.0290\n",
            "8      service                0.0272\n",
            "9         back                0.0250\n",
            "10      really                0.0239\n",
            "11        time                0.0236\n",
            "12     awesome                0.0232\n",
            "13       craft                0.0227\n",
            "14        menu                0.0222\n",
            "15      burger                0.0218\n",
            "16        cold                0.0214\n",
            "17    friendly                0.0211\n",
            "18  atmosphere                0.0199\n",
            "19       night                0.0196\n",
            "\n",
            "\n",
            "Cluster size:  475\n",
            "\n",
            "\n",
            "Cluster 5:\n",
            "          term  Average tf-idf score\n",
            "0         best                0.1656\n",
            "1         ever                0.0635\n",
            "2         food                0.0480\n",
            "3        place                0.0426\n",
            "4        louis                0.0393\n",
            "5      amazing                0.0349\n",
            "6      service                0.0321\n",
            "7        great                0.0318\n",
            "8   restaurant                0.0299\n",
            "9         good                0.0244\n",
            "10      always                0.0239\n",
            "11       staff                0.0223\n",
            "12        time                0.0221\n",
            "13        love                0.0208\n",
            "14    sandwich                0.0196\n",
            "15    favorite                0.0191\n",
            "16   delicious                0.0189\n",
            "17    friendly                0.0187\n",
            "18        back                0.0187\n",
            "19     chicken                0.0187\n",
            "\n",
            "\n",
            "Cluster size:  770\n",
            "\n",
            "\n",
            "Cluster 6:\n",
            "          term  Average tf-idf score\n",
            "0        great                0.1999\n",
            "1         food                0.0805\n",
            "2      service                0.0640\n",
            "3        place                0.0539\n",
            "4         good                0.0394\n",
            "5       always                0.0345\n",
            "6   atmosphere                0.0302\n",
            "7     friendly                0.0290\n",
            "8         love                0.0274\n",
            "9    recommend                0.0251\n",
            "10        time                0.0249\n",
            "11     amazing                0.0247\n",
            "12       staff                0.0235\n",
            "13        back                0.0234\n",
            "14     awesome                0.0231\n",
            "15        nice                0.0228\n",
            "16       drink                0.0225\n",
            "17  definitely                0.0222\n",
            "18   delicious                0.0210\n",
            "19      really                0.0200\n",
            "\n",
            "\n",
            "Cluster size:  1056\n",
            "\n",
            "\n",
            "Cluster 7:\n",
            "          term  Average tf-idf score\n",
            "0         food                0.0544\n",
            "1        place                0.0460\n",
            "2         good                0.0434\n",
            "3      service                0.0311\n",
            "4         time                0.0287\n",
            "5         love                0.0235\n",
            "6    delicious                0.0230\n",
            "7   restaurant                0.0229\n",
            "8         back                0.0226\n",
            "9       always                0.0214\n",
            "10        like                0.0212\n",
            "11      really                0.0196\n",
            "12        nice                0.0193\n",
            "13    friendly                0.0190\n",
            "14        menu                0.0187\n",
            "15       staff                0.0187\n",
            "16   excellent                0.0174\n",
            "17       would                0.0172\n",
            "18     amazing                0.0172\n",
            "19  definitely                0.0166\n",
            "\n",
            "\n",
            "Cluster size:  3126\n",
            "\n",
            "\n",
            "Cluster 8:\n",
            "       term  Average tf-idf score\n",
            "0    minute                0.0626\n",
            "1     order                0.0606\n",
            "2      food                0.0550\n",
            "3     table                0.0539\n",
            "4      time                0.0526\n",
            "5   service                0.0385\n",
            "6      told                0.0382\n",
            "7     asked                0.0375\n",
            "8      came                0.0372\n",
            "9    server                0.0366\n",
            "10    would                0.0363\n",
            "11    drink                0.0359\n",
            "12     back                0.0344\n",
            "13     said                0.0340\n",
            "14    didnt                0.0338\n",
            "15     took                0.0313\n",
            "16    never                0.0309\n",
            "17     even                0.0305\n",
            "18  ordered                0.0301\n",
            "19     hour                0.0286\n",
            "\n",
            "\n",
            "Cluster size:  1221\n",
            "\n",
            "\n"
          ]
        }
      ]
    },
    {
      "cell_type": "markdown",
      "source": [
        "### LSA Topic Modeling"
      ],
      "metadata": {
        "id": "qzDWD6d06wGT"
      }
    },
    {
      "cell_type": "code",
      "source": [
        "processed_text = df['cleaned_text'].tolist()"
      ],
      "metadata": {
        "id": "TrSesciWS0RQ"
      },
      "execution_count": 79,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "def create_gensim_lsa_model(doc_clean,number_of_topics,words):\n",
        "\n",
        "  # Creating the term dictionary of our courpus, where every unique term is assigned an index. dictionary = corpora.Dictionary(doc_clean)\n",
        "  dictionary = corpora.Dictionary(doc_clean)\n",
        "  # Converting list of documents (corpus) into Document Term Matrix using dictionary prepared above.\n",
        "  doc_term_matrix = [dictionary.doc2bow(doc) for doc in doc_clean]\n",
        "  # generate LSA model\n",
        "  #lsamodel = LsiModel(doc_term_matrix, num_topics=number_of_topics, id2word = dictionary, onepass = False, power_iters = 10)  # train model\n",
        "  lsamodel = LsiModel(doc_term_matrix, num_topics=number_of_topics, id2word = dictionary)  # train model\n",
        "  display(lsamodel.print_topics(num_topics=number_of_topics, num_words=words))\n",
        "  print('\\n')\n",
        "  index = similarities.MatrixSimilarity(lsamodel[doc_term_matrix]) # using cosine similarity\n",
        "\n",
        "  corpus_lsa = lsamodel[doc_term_matrix]\n",
        "\n",
        "  # Cluster documents by highest value in vector\n",
        "  doc_clusters_1 = []\n",
        "  for doc in corpus_lsa:\n",
        "    flattened = [val for sublist in doc for val in sublist]\n",
        "    new_list = flattened[1:][::2]\n",
        "    max_value = max(new_list, key = abs)\n",
        "    index_value = new_list.index(max_value)\n",
        "    doc_clusters_1.append(doc[index_value][0])\n",
        "\n",
        "#   df2 = pd.DataFrame(doc_clusters_1, columns = 'clusters')\n",
        "  df2 = df.iloc[:,1:2]\n",
        "  df2['clusters'] = doc_clusters_1\n",
        "  df2['sentiment'] = df['sentiment']\n",
        "  df2['stars'] = df['stars']\n",
        "  df2 = df2.sort_values(by=['clusters', 'sentiment'])\n",
        "#   #display(df)\n",
        "\n",
        "  print(df2.groupby(['clusters', 'sentiment']).count())\n",
        "  print('\\n')\n",
        "  print(df2.groupby(['clusters', 'stars']).count())\n",
        "\n",
        " # return lsamodel,dictionary,index"
      ],
      "metadata": {
        "id": "wiLZ0yQ96y2R"
      },
      "execution_count": 118,
      "outputs": []
    },
    {
      "cell_type": "code",
      "source": [
        "create_gensim_lsa_model(processed_text,9, 10)"
      ],
      "metadata": {
        "colab": {
          "base_uri": "https://localhost:8080/",
          "height": 1000
        },
        "id": "-K0xDHKNHZ_Z",
        "outputId": "00ef805d-a570-4db2-c294-2f3c55f4fd0b"
      },
      "execution_count": 119,
      "outputs": [
        {
          "output_type": "display_data",
          "data": {
            "text/plain": [
              "[(0,\n",
              "  '0.324*\"food\" + 0.295*\"good\" + 0.252*\"place\" + 0.195*\"time\" + 0.186*\"like\" + 0.182*\"great\" + 0.149*\"service\" + 0.140*\"back\" + 0.140*\"really\" + 0.138*\"would\"'),\n",
              " (1,\n",
              "  '0.570*\"food\" + -0.416*\"good\" + -0.385*\"pizza\" + -0.159*\"sauce\" + 0.146*\"service\" + -0.137*\"cheese\" + -0.123*\"really\" + 0.120*\"minute\" + -0.120*\"like\" + 0.113*\"time\"'),\n",
              " (2,\n",
              "  '0.430*\"great\" + 0.386*\"food\" + 0.310*\"good\" + 0.233*\"place\" + -0.196*\"order\" + -0.196*\"table\" + -0.187*\"pizza\" + -0.154*\"would\" + -0.150*\"time\" + -0.147*\"ordered\"'),\n",
              " (3,\n",
              "  '0.625*\"pizza\" + -0.454*\"good\" + 0.374*\"place\" + 0.286*\"great\" + -0.143*\"chicken\" + 0.084*\"time\" + -0.082*\"ordered\" + 0.079*\"crust\" + -0.078*\"really\" + -0.075*\"side\"'),\n",
              " (4,\n",
              "  '-0.482*\"pizza\" + -0.475*\"good\" + -0.399*\"food\" + 0.341*\"great\" + 0.244*\"place\" + 0.114*\"menu\" + 0.100*\"chicken\" + 0.099*\"like\" + 0.088*\"also\" + 0.082*\"sauce\"'),\n",
              " (5,\n",
              "  '0.649*\"place\" + -0.565*\"great\" + 0.212*\"like\" + -0.172*\"service\" + -0.159*\"pizza\" + -0.114*\"salad\" + -0.099*\"ordered\" + -0.089*\"cheese\" + 0.076*\"dont\" + -0.066*\"delicious\"'),\n",
              " (6,\n",
              "  '0.377*\"food\" + -0.347*\"good\" + 0.347*\"chicken\" + -0.213*\"time\" + 0.196*\"sauce\" + -0.185*\"great\" + -0.183*\"place\" + -0.178*\"table\" + -0.153*\"drink\" + 0.153*\"pizza\"'),\n",
              " (7,\n",
              "  '-0.806*\"time\" + 0.212*\"table\" + 0.173*\"place\" + 0.165*\"would\" + -0.140*\"first\" + 0.138*\"salad\" + -0.121*\"chicken\" + 0.117*\"ordered\" + -0.110*\"always\" + 0.103*\"came\"'),\n",
              " (8,\n",
              "  '0.579*\"like\" + -0.334*\"order\" + -0.308*\"place\" + -0.297*\"chicken\" + 0.277*\"really\" + -0.172*\"ordered\" + 0.152*\"food\" + -0.142*\"good\" + 0.120*\"dont\" + -0.112*\"minute\"')]"
            ]
          },
          "metadata": {}
        },
        {
          "output_type": "stream",
          "name": "stderr",
          "text": [
            "WARNING:gensim.similarities.docsim:scanning corpus to determine the number of features (consider setting `num_features` explicitly)\n"
          ]
        },
        {
          "output_type": "stream",
          "name": "stdout",
          "text": [
            "\n",
            "\n",
            "                    business_id  stars\n",
            "clusters sentiment                    \n",
            "0        0                 2760   2760\n",
            "         1                 5322   5322\n",
            "1        0                   99     99\n",
            "         1                  144    144\n",
            "2        0                   75     75\n",
            "         1                  290    290\n",
            "3        0                   81     81\n",
            "         1                  233    233\n",
            "4        0                   35     35\n",
            "         1                   77     77\n",
            "5        0                   59     59\n",
            "         1                  392    392\n",
            "6        0                   39     39\n",
            "         1                   93     93\n",
            "7        0                   58     58\n",
            "         1                  151    151\n",
            "8        0                   52     52\n",
            "         1                   40     40\n",
            "\n",
            "\n",
            "                business_id  sentiment\n",
            "clusters stars                        \n",
            "0        1.0            956        956\n",
            "         2.0            729        729\n",
            "         3.0           1075       1075\n",
            "         4.0           2087       2087\n",
            "         5.0           3235       3235\n",
            "1        1.0             48         48\n",
            "         2.0             24         24\n",
            "         3.0             27         27\n",
            "         4.0             42         42\n",
            "         5.0            102        102\n",
            "2        1.0             50         50\n",
            "         2.0             15         15\n",
            "         3.0             10         10\n",
            "         4.0             89         89\n",
            "         5.0            201        201\n",
            "3        1.0             31         31\n",
            "         2.0             20         20\n",
            "         3.0             30         30\n",
            "         4.0             87         87\n",
            "         5.0            146        146\n",
            "4        1.0              9          9\n",
            "         2.0              7          7\n",
            "         3.0             19         19\n",
            "         4.0             28         28\n",
            "         5.0             49         49\n",
            "5        1.0             27         27\n",
            "         2.0             11         11\n",
            "         3.0             21         21\n",
            "         4.0             98         98\n",
            "         5.0            294        294\n",
            "6        1.0             15         15\n",
            "         2.0             10         10\n",
            "         3.0             14         14\n",
            "         4.0             20         20\n",
            "         5.0             73         73\n",
            "7        1.0             36         36\n",
            "         2.0             14         14\n",
            "         3.0              8          8\n",
            "         4.0             38         38\n",
            "         5.0            113        113\n",
            "8        1.0             26         26\n",
            "         2.0             13         13\n",
            "         3.0             13         13\n",
            "         4.0             21         21\n",
            "         5.0             19         19\n"
          ]
        }
      ]
    }
  ]
}