metadata
title: Instacart Recommendation System
emoji: π
colorFrom: blue
colorTo: purple
sdk: docker
app_port: 7860
pinned: false
π Instacart Recommendation System
A hybrid recommendation system combining three powerful approaches:
- Item-Item Collaborative Filtering: Find similar products based on purchase patterns
- ALS Matrix Factorization: Personalized recommendations using latent factors
- LightGBM LambdaRank: Learning-to-rank with 30+ engineered features
Features
β¨ Real-time recommendations across 49K+ products π― Purchase history integration ποΈ Interactive shopping cart π Model insights and explanations
Demo Users
Try these user IDs: 1, 2, 3, 5, 8, 10, 12, 15, 20, 25
Tech Stack
- Frontend: Streamlit
- Models: LightGBM, Implicit (ALS), Custom CF
- Data: Instacart Market Basket Analysis dataset
Built with β€οΈ using Streamlit