As a way to improve my developer skills, I would like to start a project on music recommendation using ListenBrainz and MusicBrainz.
I know about music metadata and a bit about pyspark, but not enough about infrastructure, so I am not completely sure how to get started with getting access to the data to perform queries.
As a project, I was thinking of doing something like “weekly new releases”. I can see it will not be as easy as I can imagine (Fresh Releases as a playlist - #2 by rob).
I would like to use real data, not academic nor small datasets. I was playing around with the API, but I could easily reach limits, so I was wondering if there is a way for me to use pyspark in a local setup. What would you recommend me? Can I download the data dumps in my laptop and run queries on that or would that be too much for the computer?
This is our playlisting/recommendation tool – it powers LB Radio and other discovery recommendation features. It is an API-first tool, meaning does it doesn’t need to download any data, it only uses APIs. So, there is no data you need to download, most everything we have is available as an API – and if not, that API will be coming soon.
If you want to take on the weekly fresh releases – feel free to do that, but we’re not ready for that project yet. We need to make some mapping improvements before those playlists become viable.
I would suggest that you turn up in our Matrix channel and say hi here, so we can discuss in more detail: