Playlist to Podium

Playlist to Podium is a simulation that takes your Spotify playlist, analyzes its audio features, and predicts your race performance across an F1 season, turning music taste into motorsport outcomes.

Comment

Inspiration

We wanted to create a project that combines two things we're obsessed with: music and Formula 1. If playlists reflect personality, energy, and interests, could they also reflect racing ability? We wanted to see if the driver's personal pre-race playlists had any impact on their race performance, and take it a step further by giving you the opportunity to see how your own musical taste could translate to track performance.

What it does

Playlist to Podium takes in a user's Spotify playlist, extracts audio features like energy, valence, and tempo, and compares those traits to our created musical taste profiles of real F1 drivers. It then simulates how that user would perform across an F1 season — outputting race results and placements.

How we built it

We used multiple Kaggle music datasets (since Spotify's audio feature API was discontinued) and historical F1 championship data from 1950–2020. We created profiles for both users and drivers using audio features, then fed those into a multiple regression model to simulate race results and generate a season leaderboard.

Challenges we ran into

  • Spotify's audio features API was deprecated, so we had to use massive external datasets instead.
  • Cleaning and merging messy, inconsistent data across millions of tracks was tricky.
  • Getting the regression model to generate realistic, interesting race placements took a lot of tuning and experimentation.

Accomplishments that we're proud of

  • We created a full pipeline that transforms music taste into F1 race simulations — and it works!
  • We built a model that outputs meaningful, repeatable race results.
  • We found a way to merge technical machine learning with pop culture and make it really fun.

What we learned

  • How to build and tune a multiple regression model.
  • How to work with massive, real-world datasets and clean them effectively.
  • That even complex data projects can be fun, interactive, and personal when you mix in culture and creativity.

What's next for Playlist to Podium

  • Driver personality matching based on your audio vibe.
  • Dynamic race conditions: different playlists do better at different tracks.
  • Spotify integration: auto-generate your performance using real-time listening data.
  • Constructor Matching: which constructor's ethos does your playlist match best.

Built With

Share this project:

Updates