This repository includes an NLP Project about analyzing the restaurant reviews obtained from Yelp.

Problem Definition:

• What are the parameters that most affect the positive/negative evaluation of customers?
• How have the factors affecting positive/negative evaluations changed over time?

Solution Recommendation:

• Performing a sentiment analysis using the NLP method on review datas from Yelp.

Objective:

• Finding words that positively/negatively influence customer reviews.
• Comparing the results obtained from the NLP method with the comments made in 2007 with the comments made in 2017.

Webpage: https://www.yelp.com/dataset

• The Python NLP libraries, and data storages for MongoDB.

• The dataset used includes only the English language comments in 2007 and 2017.

• Performing various editing and cleaning operations on the data.

• Performing exploratory data analysis on the cleaned data.

• Performing Count Vectorizer method for detecting the most frequent words in reviews.

• In 2007, customers especially paid attention to the taste of the foods for positive voting.
• But in 2017, customers especially paid attention to the quality of customer services for positive voting.

• In 2007, customers especially paid attention to having non-smoking rooms and service time for negative voting.
• But in 2017, customers especially paid attention again to the quality of customer services for negative voting.

• For the neutral sentiment, compared to the previous 2 sentiments, clear results were not found.
• It contains both positive and negative reviews.