Qiyu Li

← Portfolio

Coursework Nov. 2024 – Dec. 2024

Spotify Songs Classification

A four-way model bake-off on 30,000 tracks, scored on accuracy, F1 and confusion structure.

What I did

  • Collected and modeled 30,000+ Spotify tracks with Naïve Bayes, KNN, Random Forest, and SVM, evaluating performance with accuracy, F1-score, and confusion matrices and finishing in the top 5% of the class competition.

Built with

  • Naïve Bayes
  • KNN
  • Random Forest
  • SVM