Tiger WangOPEN TO INTERNSHIPS
All work
ML2024

Music Prediction

A predictive model over Spotify audio features — decision trees & KNN, benchmarked with F1 scores. Findings written up as a research paper.

FIG. 04MUSIC PREDICTION04

01 — The question

Can a song's audio features predict patterns across music datasets? I used the Spotify API to find out.

02 — The models

Trained and validated decision-tree and K-nearest-neighbor models, benchmarking performance with F1 scores from confusion matrices.

03 — The paper

Authored a research paper synthesizing the findings, contributing to ongoing work in music data science and feature-based song analysis.

SPECIFICATION
ROLEResearch · Solo
YEAR2024
STACK
PythonSpotify APIWekascikit-learn
AlgorithmsDT · KNN
MetricF1
OutputPaper
NEXT PROJECTDeisBikes