i promise i'm not a weeb i just like a lot of japanese artists

spotify wrapped is bogus

2023/11/29

image: the thousand knives of ryuichi sakamoto

i woke up today at, like, 2PM and went into the kitchen. one of my flatmates was cooking spaghetti and asked almost immediately, "did you get your spotify wrapped?"

spotify wrapped is the annual summation of a given user's listening stats, their top artists, their top songs, their top genres, how many minutes they listened for, and everyone gets hyped over it like it's christmas a month before christmas. i don't really get why it happens in november and not december, since december is the actual end of the year, but i'm sure there are reasons.

this year was the first year ever, i think, where i listened to music on spotify for pretty much every month. it was my app of choice when commuting to college, shopping, sometimes even indoors when i was working. so i was looking forward to this year's "top 100 songs" playlist being varied and more interesting than last year's, but instead it sorta stank. for all 10,000 minutes i listened to (which apparently isn't a lot -- my flatmates got around 35-50,000 each), i apparently only listened to ten genres. that's 1,000 minutes per genre!

this year was huge for me, music-wise. i discovered a buttload of new artists, made a whole video about masaya matsuura's own pop band (PSY-S) in which i briefly discuss my experience listening to other japanese musicians, wrote about at LEAST a handful of albums in my media blog, and most recently i've been obsessing over the work of david byrne and talking heads. statistically speaking, i added 360 songs to my "liked songs" playlist, and certainly listened to a whole lot more via my habit of listening to entire albums at a time. i thought this was a lot, i guess it varies from person to person, whatever.

what i've got beef with is the automatically-generated playlist spotify gives every user, of their "top 100 songs" from the year. last year it just wasn't accurate because i spent up until september using youtube instead of spotify, so it was only about three months worth of listening that spotify had to work with. it made sense that it stank then, but now? it feels like it's trying to embarrass me with how samey it is. maybe i just listen to the same music a lot, but man there's a lot in there i just don't remember listening to or liking all that much. i definitely can't trust how accurate it is either!!!

according to spotify, my top five artists of the year were these:

  1. PSY-S
  2. YELLOW MAGIC ORCHESTRA
  3. CARPENTERS
  4. HARUOMI HOSONO
  5. CHELMICO

and sure, yeah, i believe i listened to that much PSY-S and YMO. CHELMICO and HOSONO don't really surprise me either, and i certainly listened to a whole lot of CARPENTERS - but there's the problem. in the automatically-generated top 100 songs list, there isn't a single song from the carpenters. they're my #3 artist! what do you mean they aren't in there!!! every other artist on this list is well-represented, so what's going on?!?

if i had to guess at all, spotify knows the genres i enjoy and leans more into those than the actual raw statistics of how frequently i listened to each song. the carpenters are an outlier in an ocean of japan's pop & rock. for the record, the top five genres were these:

  1. CLASSIC CITY POP
  2. ROCK
  3. J-ROCK
  4. SHIBUYA-KEI
  5. INDIETRONICA

and so i imagine the playlist leans heavily into these genres over softer, more melodic music like from the carpenters. i have no idea what "indietronica" is, though.

even disregarding the lack of carpenters, it's still weirdly absent of a lot of the songs i remember listening to this year. over the summer i listened to a heck of a lot of TALKING HEADS leading up to their concert film re-release, and none of that is present. when that film released, i listened to a bunch more of their music, and still nothing from them appears on the list. this is the top 100 songs, i guess the other 260 songs i heart'd include all the talking heads i listened to. the playlist doesn't even stack up with itself, though, since artists i certainly listened to heavily like RYUICHI SAKAMOTO are largely absent from the list too. he was literally in YMO! and there's loads of YMO on the list! i don't get it.

whichever mystery algorithm is responsible for generating the top 100 tracklist, i don't agree with it whatsoever. i guess it'll never be accurate, especially regarding the amount of music that simply isn't on spotify (like all of tatsuro yamashita's work). to more accurately represent the stuff i DID listen to and DID enjoy and DO remember, moderated by myself and not a faceless spotify robot DJ, here's a modified version of the playlist:

if you have spotify, and you're upset with the results of this year's wrapped, i'd recommend doing the same. you can't even share the generated list you get without transferring it into its own separate playlist anyway.

otherwise just don't use spotify. they pay artists like absolute garbage and choose to host some real shitty podcasts i won't name. also apparently the whole spotify wrapped story presentation thing was an idea stolen from an intern who wasn't compensated for it??? real scum shit.

there is also the alternative of using last.fm, a site that tracks your listening on spotify and (i think???) youtube as well as other platforms, for a more accurate summary of the year. we'll never truly have an accurate list of songs we liked this year, especially not one generated by a computer. the only way to represent what you enjoy is by doing it yourself -- i feel the same about youtube recommendations and what they reflect about us. they're never accurate!!! stop relying on computer algorithms for anything!!!