A while back, I made earbug, a little task to record my Spotify listening history using the recently played api. Amazingly, I've managed to keep it running for another whole year, so I have data to play around with.
I store the data as a zstandard compressed protobuf message containing maps, store on object storage (Google Cloud Storage). Should I maybe have used a proper database? or even sqlite? Perhaps, but this works well enough for now, and I prefer not writing SQL...
Pulling down the data, I just need to pull in the protobuf defs, and reshape the data into CSV before uploading to my analytics platform of choice: this time it's Looker Studio, formerly Data Studio.
I think I was quite consistent in the type/energy of songs I listened to over the year.
With listening (starting) times quite reflective of my schedule.
My favorite artists were no big surprise, though different from the ones that I say out loud.
While the tracks I listened to seemed to have quite the selection of vengeful(?) titles.
Next time, I might consider doing the graphing in code. Data/Looker Studio isn't as flexible as I want in terms of reshaping the data and adding some more interesting derived data points.