machine learningaudiodata scienceschool2025

SoundScope

Completed

Overview

SoundScope is an interactive music analysis tool that blends Spotify-style audio features with machine learning predictions inside a Streamlit interface. Built as a machine learning class project, it functions as both a data science demo and a fully interactive app, letting users explore songs, visualize audio characteristics, and experiment with a trained mood-classification model.

Why It Exists

Built for a machine learning course. The goal was to make ML tangible: not just a notebook that spits out accuracy numbers, but something you could actually interact with and see working on real music data.

Features

  • ·Curated static dataset of well-known tracks with audio feature approximations
  • ·Mood classification model trained on audio features (valence, energy, tempo, etc.)
  • ·Interactive Streamlit app for song exploration and feature visualization
  • ·Audio feature breakdowns with charts and comparisons across tracks
  • ·Training pipeline with Jupyter notebooks documenting the full ML process

Tech Stack

PythonStreamlitscikit-learnpandasJupyter

Links