Recommender system for
restaurant menu app

Project Description:

As part of the project, we’ve scraped data about meals and menus from the internet. More than 7 million menu items from around 40,000 restaurants in the USA.

Feature extraction techniques were implemented in order to transform semi-structured data into a more useful representation. A personalized content-based menu recommendation system was created.

We’ve incorporated filtering techniques, specifically to deal with the cold start problem.

Technologies:

• spaCy, PyTorch.
• Scrapy framework to collect data from the internet.