Inspiration

We built our tag.bio platform to answer deep questions in any dataset. But we hadn't tried it out yet on a dataset with music and track/metadata - so we wanted to see what we and our platform could achieve in 24 hours of this hackathon.

What it does

Using data (800k tracks, 5k "most popular" artists) pulled from the Musimap API, we can now compare and contrast different styles and attributes of musical tracks. Ideally, when combined with user metadata and listening/rating behavior, we could unlock valuable new temporal, regional and demographic trends in listening behavior.

How we built it

For the hackathon, we wrote code to pull as much data as we could from the Musimap API, then encoded 3 modules into the tag.bio platform for specific analysis of that dataset.

Challenges we ran into

As with most hackathons, getting the API connector to work properly is half of the battle. We weren't able to get all the data from the API (that would take a week with current download speeds), but we got enough tracks for a functional demo.

Accomplishments that we're proud of

We created analysis modules that identify track metadata correlations for any term or artist within 30 seconds. Anyone that goes to https://ardentes.tag.bio can find their own interesting results via point-and-click.

What we learned

This industry is transforming with data, and especially with new regulation like GDPR. It will be important to extract as much value in the future, while preserving user rights to data collection and use.

What's next for ardentes.tag.bio

More data, please! And more questions - the best analysis modules are created by an expert in the industry who has some important questions to ask.

Share this project:

Updates