scorecard
  1. Home
  2. entertainment
  3. Music
  4. Spotify is learning from your listening habits to give you more music you'll actually like

Spotify is learning from your listening habits to give you more music you'll actually like

Alex Heath   

Spotify is learning from your listening habits to give you more music you'll actually like

Spotify

Reuters

When I'm at work, I listen to a lot of instrumental, ambient music. I usually fire up Spotify on my Mac and play a playlist like "Deep Focus," on shuffle for hours.

I'm not actively listening to anything in particular - I just want something non-distracting playing in the background.

It's a different story when I'm out and about or even at home. I skip heavily through my favorite songs at the time and sometimes listen to a particular track I'm really digging on repeat. I also like to be more adventurous and listen to artists I'm not familiar with on Spotify.

I felt like I knew my own listening habits. Then I saw how Spotify sees what I listen to.

Every Spotify user has what the company internally refers to as a "taste profile." It's a collection of everything you listen to on Spotify broken down into genres and subgenres, like chillwave and alternative hip hop.

The goal of your taste profile is to "come up with a nuanced understanding of each portion of your taste," Ajay Kalia, who oversees the project, tells Tech Insider.

"We believe that it's important to recognize that a single music listener is usually many listeners, and a person's preference will vary by the type of music, by their current activity, by the time of day, and so on," he says.

I have two main kinds of music of I listen to in Spotify that are grouped into what are called "clusters."

screen shot 2015 08 24 at 5.52.40 pm

Spotify / The Echo Nest

How Spotify sees my taste in music

My largest cluster is 98 artists who are primarily hip hop-oriented, like Kanye West, Kendrick Lamar, Drake, Miguel, and Jay­ Z. My second, smaller cluster has about 20 artists that are primarily instrumental, like This Will Destroy You, Explosions in the Sky, and Caspian.

Within those main clusters, Spotify knows the level of affinity I have for certain artists, explains Kalia. Based on how I listen, Spotify makes "inferences about the artists you enjoy, the artists you don't, and the artists where we haven't learned enough to know," he says.

Here are the criteria for how Spotify determines your affinity score with an artist:

  • How heavily you play
  • How regularly you play
  • How "actively" you play (such as playing from your playlists or an album vs. a radio station)
  • How much you play from an artist's full catalog of music

For instance, I have a very high affinity in Spotify with Ben Howard, but a low affinity with The Black Eyed Peas - which is true. While I've played both artists on the service before, Spotify knows that I like one and don't particularly care for the other.

The system allows Spotify to edge out actively that doesn't fit the overall pattern of your listening habits. Taste profiles are the foundation of personalization at Spotify, says Kalia. They aid the company in making new features like Discover Weekly, a playlist for every Spotify user that's updated once a week with new music based on what they listen to.

At the end of the day, explains Kalia, Spotify is trying to get your music taste right.

NOW WATCH: How to use Apple's Spotify killer - now on everyone's iPhone

READ MORE ARTICLES ON



Popular Right Now



Advertisement