Restaurant Review - Analytics

Motivation

There is a vast amount of reviews based on user likes and tastes which needs to be analyzed to recognize user sentiment and develop language based models for review generation and recommendation

Objectives

1. Analyze the reviews for restaurants distributions, spatial mapping and statistics
2. Develop a review sentiment analysis model for ordinal regression of rating
3. Develop a review generation model based on Natural Language Generation
4. Develop a recommender system based on the review

Data

• Restaurant data is acquired on census block groups based on high foot traffic data

• Reviews are scraped from maps and only 80% of reviews are taken where total reviews > 500

• The review scraper was based out of Gaspa93's repo

• Restaurants with 'Hotel' , 'Inn', 'Plaza' or 'Intercontinental' in their name were omitted due to scraping errors