Facing the music
First, I had to figure out an efficient way to rate each album. I devised my own formula (more details at my GitHub repository) to produce an album rating on a 100-point scale. With the formula set, I started listening to all the music, transferred the ratings to a CSV file, and plotted the results. Using my results as a guide, I've purchased eight albums already (and plucked one in this list for free from a family house cleanout).
Lining up the acts
With so many ratings for individual albums, I had plenty of data to play with to look at other aspects of my musical preferences. My first instinct was to try to devise a way to rank my favorite artists. After many iterations, I found another formula that not only seemed to be pretty fair but also stacked up my favorites like I had hand-picked the list myself. Note that artists were only included once I listened to their full discography (that is, all studio albums, for simplicity). Essentially, the artist-ranking formula tries to take into account the quality of an artist's full discography and individual songs. An artist's minimum, median, and maximum album rating contribute to a final ranking score in addition to a normalized mesaure of how many outstanding (or "starred", in this case) tracks I think an artist has made. (Again, for details on the full formula, visit the repository.)
Mixing the rest
Once I finished laboring over the creation of two formulas, I was free to look at more straightforward items. Below are six more charts that each examine a different aspect of my data.
I've listened to at least one album recorded in every year since 1962, but I have listened to a lot more albums that have been recorded in my lifetime (beginning circa the 1990s).
Was I likelier to rate an album more highly based on the year it was released? Once the small sample sizes of the 1950s and 2020s are factored out, my favorite decades appear to be the 2010s, 2000s, and the 1970s. Not too high on the '90s, though.
I didn't join a music streaming service until 2016. Joining Spotify clearly allowed me to massively increase my music consumption — the number of albums per year I listened to for the first time ever ramped up starting in 2016. Before that, I relied on my own purchases (CDs or iTunes) and shared music from friends (files or burned CDs).
Some analysis has shown that our strongest musical preferences set in around age 13 (for women) or 14 (for men). Did the year in which I first heard an album (recorded at any time) mean I was likelier to give it a higher rating? It seems so, but the earlier years also benefit from containing a significantly lower number of albums than the last few years.
I tend to have a more positive reaction when seeing an album is a nice, tight 40 or so minutes long before listening to it. Did my ratings reflect this preference? It looks like it. Make an album between 30 and 70 minutes long, and it's possible I'll be happier with it.
Additionally, I wanted to see if there is a specific track count that I prefer an album to have. Eight or nine is the sweet spot, but they and the next top two track lengths (14 and 17) might have an advantage from having fewer albums than the middle options (10 to 13).
Running out the groove
What I enjoy about this endeavor is that it's a living project. I'll never be done with it; I get to keep adding data. As I continue to listen to music, I'll keep this page updated, and I'll note any changes to the data. Thanks for taking a look!