Music Recommendation System

Overview

This project builds a music recommendation system by analysing audio features from music datasets. It leverages clustering and similarity algorithms to group tracks by audio characteristics such as tempo, energy, danceability, and valence. Given a track, the system recommends songs with similar sonic profiles — replicating the core mechanism of streaming platform discovery engines.

Key Features

• Audio feature extraction — tempo, energy, danceability, valence, and more
• K-Means clustering to group similar-sounding tracks
• Cosine similarity for song-to-song recommendation scoring
• Exploratory data analysis with Pandas and Matplotlib
• Dataset visualisations showing feature distributions per cluster