Movie Recommendation Using Semantic Similarity of Plot


Search for movies by title or describe the plot, and we’ll find you films with a similar story—your next favorite movie is just a query away!

Project Details

  1. Data Storage: Movie data is stored in MongoDB, including features such as titles, genres, descriptions, and vector embeddings.
  2. Generating Embeddings: Using the OpenAI API, we generate vector embeddings for movie descriptions.
  3. Vector Search: MongoDB’s vector search capabilities are used to find similar movie embeddings, providing recommendations based on user input.
  4. User Interface: Streamlit is used to build a web app where users can enter their preferences and receive personalized movie recommendations.

Screenshot

Retrived results based on the following plot:

Earth isn’t inhabitable anymore and a group of brave astronauts must undergo a brave mission of inter-galactic travel to save the humanity from extinction.

screen shot of the app

Written by

Fran

Tariq Mehmood

Machine Learning
Engineer