sdl60660/letterboxd_recommendations

Scraping publicly-accessible Letterboxd data and creating a movie recommendation model with it that can generate recommendations when provided with a Letterboxd username

What it solves

This project provides a movie recommendation system that suggests films based on a user's Letterboxd profile. It eliminates the need for manual searching by analyzing a user's existing ratings to predict which other movies they might enjoy.

How it works

The system scrapes public star ratings from a Letterboxd profile and converts them into numerical values (1-10). It then uses a collaborative filtering model based on Singular Value Decomposition (SVD) trained on a large sample of ratings from other users. The model predicts scores for movies the user hasn't seen, returning the top-rated predictions. It is entirely based on rating patterns and does not use movie metadata like genre or cast.

Who it’s for

Movie enthusiasts who use Letterboxd and want personalized recommendations based on their rating history.

Highlights

  • Collaborative Filtering: Uses SVD to find similarities between users and movies.
  • Full-Stack Implementation: Includes a React frontend, FastAPI backend, and MongoDB for data storage.
  • Full-Cycle Data Pipeline: Features custom crawlers to scrape Letterboxd data and Redis for task queue management.
  • Large Scale Data: Capable of processing millions of ratings to generate predictions.

Related

  • Project
    klaudiosinani/moviebox

    A content-based movie recommendation system that uses TF-IDF and cosine similarity to suggest movies with similar plot summaries.

  • Project
    enoche/MMRec

    A modern multimodal recommendation toolbox that provides a unified framework and source code for a wide variety of research-based recommendation models.

  • Project
    recommenders-team/recommenders

    A comprehensive library of recommendation system algorithms and best practices, providing tools for data preparation, model training, evaluation, and production deployment.

  • Project
    MTSWebServices/RecTools

    A Python library that streamlines the creation of recommender systems by providing a unified interface for various models, from matrix factorization to advanced Transformers.

  • Project
    interviewstreet/hiring-agent

    An AI-powered resume-to-score pipeline that extracts structured data from PDFs and GitHub profiles to rank candidates based on explainable, role-specific rubrics.