Better Spotify Wrapped

Every year Spotify shows me a summary of my music listening for the year, but I'm never impressed because I feel like I didn't actually learn anything. Let's see what we can actually learn from my listening habits.

Genre Space

Let's start with the coolest chart: a map of my top genres compared to everyone else's. Hover over each dot to see what that point represents.

This map is built with data provided by Glenn Mcdonald's project Every Noise at Once, where he identified thousands of individual genres and their properties in nearly a dozen dimensions. I'm showing five dimensions here:

• Density vs Bounciness along the x-axis (left to right, respectively)
• Organic vs Mechanical along the y-axis (down to up, respectively)
• Energy, red color channel
• Dynamic Variation, green color channel
• Instrumentalness, blue color channel

Music is not distributed along all of these axes equally. The background is a density map of genres in the ENAO database, showing where on the map more artists tend to congregate. There's a rough line from the top-right (house, techno, EDM) to bottom-left (choir, folk, instrumental) with a bulge near the middle representing the rock and pop regions. I've labeled a few of those regions with their dominant genres.

So where do my tastes tend to lie within this map? I'm almost entirely in the northern, more mechanical half of this chart - my southernmost genres include a few indie and instrumental bands but nothing further than that. I have a few genres I like along the very north, like pop punk to the west and house to the east. The bulk of my hours, however, are spent solidly within the rock/pop region.

Breaking these out, you can see that I have a large preference for genres with a high dynamic variation and, to a lesser extent, high energy and mechanicality. I think these all match very closely with how I would describe my tastes - I get bored with music easily and it needs to "tickle my brain" to keep me interested.

Now how have these genres shifted over time?

Here, I'll let you scrub through my entire musical history. I've isolated the top genres for every month of my listening and you can show each of them by dragging the slider.

You can see the early days where I mainly listened to pop punk (Panic!) and slowly migrated to a more varied pop and indie palette. You can also see the months where I listened to a bunch of jazz (December 2021) or vocaloid (July 2024). In the end, though I'm still centered firmly in that rock/pop region.

For even more proof that my tastes don't really change, here are the centroids of my genre distributions for each month with the routes between them.

It's like the tracks after an ice skater does a spin right around (38, 68).

Listening History

Spotify gives me some yearly stats, but what about a monthly breakdown? How long did I listen, and what were my top songs and artists every month?

Let's plot the number of listened hours over time.

This plot was very surprising to me - I thought I was listening to a lot of music recently, but apparently I was listening way more in 2019. For a few months during the summer of 2019 I was listening over 8 hours per day, essentially non-stop from when I woke up to when I got home after work.

How picky am I about what plays? This chart shows the total duration of song I've ever skipped, which is quite a lot. In fact, there were even a few months where I skipped more music than I actually played. It seems I just wasn’t feeling whatever Spotify was recommending during that time.

I don't listen to a ton of explicit music, but I'd be very curious to see how this stacks up against other users or what proportion of songs on the platform are marked explicit. Apparently my tastes have gotten only slightly edgier over time.

I used to hate shuffle, and I never ever used it. Now I use it all the time. Pretty much the only time I don't shuffle is for an album or playlist someone has put in order on purpose.

Here I've separated out each play into morning (4AM-12PM), afternoon (12PM-8PM), and night (8PM-4AM). Most of my listening nowadays happens sporadically through the workday or evening, so this distribution rings true.

This chart is shaming me very thoroughly - I've been doing a decent job at finding music new to me, but that music is still pretty old on average. I should work on finding some music that's a little more current, or looking into the albums that my favorite artists have released recently.

Artist Diversity

Do I have a couple favorite artists or do I listen to a lot equally?

Well it's certainly not equal. I have just a couple artists I listen to a lot and a lot that I've only listened to once or twice. Just got more than 20 hours of stream time, but I've listened to artists total! This actually does seem like a good thing overall for my exploration - listening to a lot of artists just a few times is a great way to find new favorites.

How deeply do I explore each artist's discography? If I like a song, do I go explore that artist's albums?

Well that's also a pretty clear "no". There's only artists where I've listened to over a dozen songs.

Now if we combine these two attributes, we can find which artists I've listed to a lot but not explored thoroughly. Any artists in the bottom-right strip would good to listen to more of.

Can we calculate the best artists to explore algorithmically? Sure can. I would simply divide each artist's number of hours played by the number of songs I've listened to, but that would just show a lot of artists I've listened to once and don't have any particular interest in hearing more of.

Instead, I shift each value a little before calculating this score. I've found that shifting the duration by 3-5 hours helps remove the artists I don't really care about. I don't think shifting the play count is very helpful but I've included a slider here so you can try it for yourself. You can imagine this score as the slope between this point and each artist's point, so I've plotted it on the graph above and highlighted the top-scoring artists.

The last thing I do here is filter out artists I've already listened to a lot (both total duration and number of songs). This is intended for me to find hidden gems, not listen to GROUPLOVE for the millionth time. Here's the final list:

Popularity

How closely do my tastes match everyone else's?

Let's plot the global popularity along the x-axis, and how much I listen to each artist on the y-axis. If my tastes match everyone else's, you'll find a straight line from the bottom-left corner to the top-right.

Well it's certainly not a straight line, except maybe a vertical one. Let's stick the x-axis on a log plot, since it seems it may be scored that way. (The Dragons have skewed my plot once more.)

Here we can actually see something! My favorite bands to already be reasonably popular, above Spotify's popularity score of 70,000. I do still have a fair number of bands I like below that point (hipster cred) but I don't listen to them as much.

Now let's do the same thing for individual songs:

My tastes in individual songs are much more evenly-spread. I've listened to super-popular songs just as much as some that basically nobody has ever heard of, according to Spotify. Though I do doubt some of these - are you telling me that I'm the only one listening to Australia by The Shins?

Miscellaneous

For fun, let's look at some other random stats.

Song Durations

Is there a significant difference in the durations of songs per genre?

If there is an affect, it is very slight. Neo-synthpop is the longest genre on average, while "alt z" is the shortest, but the spreads on both are wide enough that I don't feel comfortable making any conclusions about this.

Discoveries

Every now and then, you come across a song that gets lodged in your head. To track those, I've compiled a list of the most-played songs each month that were played for the first time that month (with a minimum of three plays). Reading through the list was a fun trip down memory lane, since several of these were songs I played a lot for a short time but never added to a playlist I listen to now.

Often-Skipped Artists

Are there artists that I listen to a lot, that I also skip a lot?

Very much yes. Many of my favorite bands get skipped a lot! I think this is a combination of me not being in the mood for everything all of the time and Spotify shuffle bringing up certain artists way more than it should. Anything in the top-left section here is music that Spotify thinks I like way more than I actually do (Mother Mother).

Genre Artists

For a nice deep dive into my favorites, here's all of my top genres and my top-played artists in each one.