Remember scrobbling?

Ever even heard of scrobbling?

Anyway. For the last dozen or so years I’ve configured my primary music player (current MediaMonkey) to send last-played data for songs in my library to the website. The idea being that folks can see what I’m currently playing and/or most recently listened to. That’s entertaining in and of itself, at least to me. There’s a bonus, however. Doing big-data things to my music-listening information paired with similar information from other users results in an ability to recommend new music. The idea goes something like this:

  1. I listen to These Songs by These Bands quite a lot.
  2. A stranger on the Internet also listens to These Same Songs by These Same Bands, a lot.
  3. This stranger also listens to Some Other Songs by Some Other Bands.
  4. In theory, there’s a good chance that I might also like those Other Songs and/or Other Bands based on a commonality of musical tastes with a stranger on the Internet.

It’s a great idea. I even found some new-to-me musical artists as a result from time to time, such as Way Out West.

However. has… degraded somewhat in usefulness over the course of this decade, and when DJ Sundog over on Mastodon recently noted that a self-hosted alternative exists, I had to try it out. Let’s be clear that I’m basically losing the “big data” part of what made (and theoretically, its erstwhile replacement) valuable. Now I just want somewhere to point my data to. How I’m going to use that data is a problem for another day. I love building things!

Here’s what I’ve learned. (And, yes, I need to write about my Mastodon instance at some point as well. Please be patient; I’m out of the habit of actually blogging.)

Continue reading