What I've learned in ten years tracking my listening on Last.fm

by

A partial analysis of Jay Gabler's 2016 listening
A partial analysis of Jay Gabler's 2016 listening (Last.fm)

On Jan. 5, 2007 at 9:10 a.m., I listened to R.E.M.'s song "Bubblegum Pop." Little did I know that it would become the first of 105,193 song listens that I would track over the next decade via the social network Last.fm.

I'd signed up for Last.fm at the suggestion of my friend Alice, who pointed out that it was interesting to compare music tastes. At first I curiously paged through the profiles of my friends like Alice, to gauge our compatibility. (My compatibility with Alice is "very high," Last.fm deems.)

I was also fascinated at the ability to find the people on the site who were most musically compatible with me. The UK-based network cutely calls these "neighbours." My musical neighbours currently include a user named Helen (who's also been grooving to Solange and Shura lately), a person named Nevid whose profile picture is Angel Olsen in her golden wig, and fellow Mitski fan Niko.

As time went by, though, I lost interest in the social aspects of the site. What became increasingly fascinating were my personal charts: the tabulation of my complete listening history since I started "scrobbling" on the site. An app tracked my iTunes listening, and when Spotify came along, an in-app integration meant that all of my Spotify listening was also tallied on Last.fm.

When someone asks what my favorite artists are, I can now answer with great precision, at least as regards my digital on-demand listening since January 2007. The artists I cue up most often are Bob Dylan (3,278 song plays), Tegan and Sara (2,324 plays), and the Magnetic Fields (2,047 plays). My top tracks of the decade were Kate Nash's "Do-Wah-Doo" (66 plays), Rihanna's "We Found Love" (64 plays), and Tegan and Sara's "Closer" (61 plays).

The site even lets you break your listening down by specific time period — in case I'm curious about what, say, my favorite song of 2011 was ("Video Games," Lana Del Rey) or what album I grooved to most often in 2016 (the deluxe edition of Fleetwood Mac's Tusk). Is it self-absorbed to dig through all of this? Of course it is, but I think every music nerd can identify with the fascination at least a little bit — and so can every data nerd.

Looking back on a decade of listening, here are a few additional insights I gleaned.

Prince is right: Albums still matter

For me, at least. When I look at the artists I've heard the most tracks by over the past ten years, it's the ones with broad and deep album catalogs that rise to the top. Dylan, R.E.M., Springsteen: these are artists with an album, and sometimes several, for every mood. Over time, those hours add up.

My listening is oriented towards music discovery

That's not a brag, it's a personal reassurance that I'm not terrible at my job (working at a music station) just because it seems like every year is full of amazing music that I don't get around to hearing. According to Last.fm's year-in-music analysis, 68% of the tracks I listened to in 2016 were tracks I'd never heard in 2015 or prior. For my closest "neighbours," that rate is 50% or less.

That said, I'm not as interesting as I think I am

When someone asks me what music I like, of course I'd like to answer something along the lines of, "Very eclectic! A little bit of everything! You can't box me in! I embrace all genres!" The top tags among my most-listened artists, however, don't lie: "indie," "rock," "pop," "classic rock." Gee, could you ever possibly guess I'm a 41-year-old white guy from Minneapolis?

The music that means the most to you is not necessarily the music you listen to most often

This was an insight I reached after five years on Last.fm, when I wrote about my listening history for The Tangential. I don't think I could put it any better today.

"I've listened to Lana Del Rey's Born to Die so many times that she landed among my top ten most-listened artists," I wrote, "but it's not an album that means a lot to me emotionally, it's just nice to listen to while I work. Some songs I treasure are among my most-listened — Tilly and the Wall's 'Rainbows in the Dark,' Jenny Lewis's 'You Are What You Love,' Frightened Rabbit's 'I Feel Better' — but others, like the recording of 'Thunder Road' that opens Springsteen's live box set, are tracks I come back to only occasionally, when I really need them. There are people in my life like that, too."

When it comes to recommendations, there's no substitute for the human touch

If you knew every song I'd listened to over the past ten years, how well would you do at recommending my next track? Not bad, it turns out — even if you're a robot. Last.fm offers recommendations, which currently include "Antabus" by Makthaverskan (never heard of it) and "Won't Back Down" by Tom Petty (an old favorite). With increasingly comprehensive data on listening habits, services like Last.fm and Spotify have become quite sophisticated about creating listening recommendations that balance the typical user's desire for a balance between familiar tunes and complementary new discoveries.

Though I do find these tools useful, robots can only do so much. Whether or not it's actually recorded on a tape, a personalized mixtape just hits you in a way that an automatic playlist can't — and even if it's not created personally for you, a playlist like those my friend Abbie makes has the cohesion and fascination of a set of songs curated by a real live human being. That's what our hosts bring to The Current, because after all, music is best when it's shared.


comments powered by Disqus