Restaurant Reviews
Database Management and Sentiment Analysis
To cope with all the sleepless nights in pursuit of finishing my gazillion submissions, I rely on caffeine. NO, it doesn’t help one bit! I think it’s all in my head. But atleast all my Starbucks trips gave me the idea to take up this project. The aim of this project is to analyze the reviews and tips given by customers to evaluate Starbucks Coffeehouses in and around College Park area.
![](https://framerusercontent.com/images/7oFM7crbrQQPku8gFHGlY9GVP54.jpg)
--
For this project, I started off by scrapping the data from the following websites:
Yelp!
Google Reviews
FourSquare
The data consisted of datapoints such as customer name, review, feedback, check-in time, business name, business address, etc.
After getting the data in place, I made an ER diagram using Lucidchart. This ER diagram further helped me to correlate things of my interest. I also defined certain business rules, made a relational model, and implemented normalization.
![](https://framerusercontent.com/images/cYfsk4flP3YIK6WzwXcxc6YVjY.png)
The next task was to create tables and insert corresponding data. I saved these SQL queries in a DDL file. After inserting data into the tables, I made a DML file to create materialized viewsusing complex SQL queries in order to answer the following questions of interest:
Which is the highest rated Starbucks branch in College Park area?
What sets the highest rated branch apart from the rest?
How can we improve the lowest rated branch?
How can we improve staff allocation strategy?
After getting answers to the above mentioned questions, I decided to do a Sentiment Analysis of customer reviews to better understand their opinions.
Moreover, I made a Tableau dashboard in order to visualize the results and compare the branches easily.
![](https://framerusercontent.com/images/9ssJgPn8Ks3mUmpbopAn1gVwRA.png)
I believe that this project can be used by any business in order to refine their understanding of customers and translate into improved sales.