Inspiration

We were inspired by the poor recommendations of the Spotify Weekly algorithm to create our own solution.

What it does

You can sign into Discover Weekly 2.0 with your Spotify account and it will pull your top 50 songs, plotting them along a scatterplot. You can then plot new songs, and view similar recommended songs.

How we built it

We used a Jupyter notebook to create the back end, and the Express Node.js framework to create the front-end. We also leveraged the Spotify API for account syncing.

Challenges we ran into

Integrating the Plotly library with the Express Node.js framework to dynamically create scatterplots was challenging, as well as understanding how to fetch data using our third-party API.

Accomplishments that we're proud of

We are proud of the ability to embed songs using ML algorithms, and to leverage thousands of datapoints from the Spotify API.

What we learned

We learned how to implement an API within a web framework, particularly as neither of us have front-end experience. We also learned how to implement a front-end in the Express framework with Node.js.

What's next for Discover Weekly 2.0

We hope to add functionalities to sync multiple Spotify accounts and plot their songs against each other. We also hope to extend our functionality to an app framework.

Built With

Share this project:

Updates

Submission history