Yelp-Restaurant-Reviews-NLP-Sentiment-Analysis

This project uses Yelp's busienss and review datasets to conduct analysis on customers' reviews on different types of cuisines in the United States.

The Yelp dataset is downloaded from Kaggle website. In total, there are 5,200,000 user reviews, information on 174,000 business. we will focus on two tables which are business table and review table.

Dataset Link: https://www.kaggle.com/yelp-dataset/yelp-dataset

Table of Content:

1. Overall Project Objectives

2. Description of Data

3. Clean Yelp_business dataset

4. Clean yelp_review dataset

4.1 Merge two datasets and get new dataframe restaurants_reviews

5. Exploratory Data Analysis

5.1 Restaurants Distribution- Types of restaurants

5.2 Top 10 cities with most restaurants

5.3 Distribution of restaurants in each State

5.4 Restaurant Reviews Distribution

5.5 Top 10 cities' restaurants received the most reviews

5.6 Top 10 restaurants received most reviews

5.7 Distribution of positive and negative reviews in each restaurant category

5.8 Avergae Rating of Each Type of Restaruant

5.9 Average length of reviews by positive/negative reviews

5.10 Average Review Length by Restaurantes Types

5.11 Ratings Distribution

6. Cleaning and Processing Text Data

6.1 Top 10 unigram

6.2 Top 10 Bi-grams

7. Naive Bayesian Model

8. Logitstic Regression

9. Conclusion