Eating at a restaurant with a group of friends and not liking the food at all is the worst feeling - you went to the restaurant because a few people in the group said it was a good place to go to and now you're left sitting there awkwardly as the others enjoy the food and you just cannot. We wanted to create a regression model based on a dataset of past reviews and past meals we'd eaten to successfully determine an optimal restaurant to go to for a group of people and eat easy.

What it does

eatsy is a regression model that models what preferences are important to individual users and extrapolates this data to recommend a restaurant that satisfies the preferences for a group of users. It takes into account reviews on other online platforms and user-specific preferences and past experiences to find the perfect restaurant for everyone.


  • Models with all the data in Yelp's open dataset that holds information on 200,000 businesses, over 6 million reviews, and 1.6 million users
  • Implements machine learning techniques to predict the importance of certain preferences to a user
  • Combines the importance of preferences of every user in a group optimally
  • Uses a logistic regression model to calculate a score for each restaurant given a group of users
  • Selects the top five restaurants that produce the highest scores for the group of users

How I built it

Logistic regression with Sklearn, Extracting data from Yelp dataset with Json, Flask to present findings.

Challenges I ran into

Parsing through large amounts of data to train the regression model.

Accomplishments that I'm proud of

Building an in-depth regression model that resourcefully uses Yelp's open dataset. Learning new tools, and producing results that people can use on a daily basis.

What's next for eatsy

Incorporating regression analysis using categorical variables.

Built With

Share this project: