Skip to content

lorenzenv/Movie-Recommender

Repository files navigation

Movie Recommender

KMeans, Flask, Heroku, theFuzz and Sprase Matrices

This project was completed in week 10 of the Spiced Data Science Bootcamp in Berlin. The goal was to create a Movie Recommender Webservice running on a Heroku Server.

The front-end was build with Flask and Bootstrap. It uses thefuzz and tmdb libraries to pull movie posters from the TMDB api.

Predictions are made with a KMeans Model, trained on a database of movies and ratings.

The model finds 10 similar users and shows their top five rated movies.