Restaurant Rating

Inspiration

Simple averages are not a sophisticated enough way to assess online reviews considering some people are more trusted reviewers than others. We wanted to create a better way to assess online reviews.

What it does

Takes Yelp reviews. Finds the figures out the quality of each reviewer, and the similarity between each review and other reviews. Each review is given "importance points" based on its quality which are then sent to other similar reviews in a steady transition system. This process continues reaching a steady state.

How we built it

A different team scraped the data from Yelp and sent it to us. We then used pandas python and data science library and used an algorithm to assess the quality of each review. We then processed the data to model the steady-state transition system.

Challenges we ran into

We had trouble integrating the two main components of our project (quality assessment, steady state) and ultimately couldn't quite do it. So we used simulated data to test our steady state.

Accomplishments that we're proud of

We're proud to apply the concepts we learned in this and other classes such as Pandas, Colab, data science lifecycle, and linear algebra to create our very own proprietary algorithms.

What we learned

-Work with matrices and tables -How different statistical distributions can be used to simulate data -How to use data science techniques like standardization and one-hot encoding