All work

Machine Learning / Content-based recommendations

Movie Recommendation System

A content-based recommendation system that finds similar movies from metadata and fetches posters through TMDB.

  • Python
  • Pandas
  • NumPy
  • Scikit-learn
  • NLTK
  • Streamlit
  • Requests
  • TMDB API
Try LiveGitHub

Approach

The system combines movie metadata into tags, applies Porter stemming and CountVectorizer, ranks items with cosine similarity, and uses TMDB for poster fetching.

Architecture

01Movie metadata
02Tags from genres, keywords, overview, cast, and director
03CountVectorizer
04Cosine similarity
05Top 5 recommendations
06TMDB poster fetching

Capabilities

  • Content-based recommendation
  • Porter stemming
  • CountVectorizer
  • Cosine similarity
  • Top 5 similar movies
  • Streamlit UI
  • TMDB poster fetching
  • Pickle-based processed data