A machine learning–based song recommendation web app built with Streamlit, PostgreSQL (Supabase), and K-Means clustering.
Users can search for songs, select a track, and get recommended tracks with similar features based on cosine similarity.
- 🔍 Search songs by title or artist name
- 🎧 Get top-N recommendations based on track similarity
- 📊 Interactive table and similarity bar chart
- 🖼️ Displays album artwork and audio preview (if available)
- 💾 Download results as CSV
- ☁️ Deployed using Streamlit Cloud or AWS EC2
| Component | Technology |
|---|---|
| Frontend | Streamlit |
| Backend | Python, SQLAlchemy |
| Database | PostgreSQL (Supabase) |
| ML Model | Scikit-learn (KMeans, Cosine Similarity) |
| Deployment | Streamlit Cloud / AWS EC2 |
| Data | Spotify audio features dataset |
- 🖥️ AWS EC2
- URL: http://13.49.134.126:8501
- (Instance may be inactive — please contact me to activate it)
- ☁️ Streamlit Cloud