Become a Member
Remove ads, unlock a dark mode theme, and get other perks by upgrading your account. Experience the website the way it's meant to be.

Spotify’s Discover Weekly: How Machine Learning Finds Your New Music

Discussion in 'Article Discussion' started by Melody Bot, Oct 16, 2017.

  1. Melody Bot Oct 16, 2017
    Melody Bot

    Your friendly little forum bot. Staff Member

    This article has been imported from chorus.fm for discussion. All of the forum rules still apply.

    Sophia Ciocca, writing for Hacker Noon:


    The exact mechanisms behind NLP are beyond the scope of this article, but here’s what happens on a very high level: Spotify crawls the web constantly looking for blog posts and other written texts about music, and figures out what people are saying about specific artists and songs — what adjectives and language is frequently used about those songs, and which other artists and songs are also discussed alongside them.

    While I don’t know the specifics of how Spotify chooses to then process their scraped data, I can give you an understanding of how the Echo Nest used to work with them. They would bucket them up into what they call “cultural vectors” or “top terms.” Each artist and song had thousands of daily-changing top terms. Each term had a weight associated, which reveals how important the description is (roughly, the probability that someone will describe music as that term.)
    ​

    Expand - View Original
     
  2. ReginaPhilange Oct 16, 2017
    ReginaPhilange

    Trusted Prestigious

    I love the playlist it makes, I've found some great stuff from it. also:


    3 part series on neural networks for anyone interested. Still waiting on the third video. I recommend his videos in general, he does a great job of visually explaining these topics.
     
    mercury likes this.
  3. Shrek Oct 17, 2017
    Shrek

    can't be made fun of Prestigious

    this is very cool. i'm a broken record in these threads but their personalized playlist curation is the killer app in streaming music for me. the interface could be garbage forever and i'd put up with it for all the stuff i find and enjoy that i'd other wise never listen to.
     
  4. Ferrari333SP Oct 17, 2017
    Ferrari333SP

    Prestigious Supporter

    I've a heavy Spotify user, and have been listening to Discover Weekly for a long time, and I'll probably get 1 track out of like 5 Discover Weekly playlists that I actually like, on average. Compared to the people who written about Discover Weekly, I almost never find anything I like on the list
     
  5. Remove
  6. personalmaps Oct 18, 2017
    personalmaps

    citrus & cinnamon Prestigious

    having worked with the NLP team at my company here and there, it's so cool to see how this is applied in other industries. I don't have the most comprehensive understanding of it, but I love trying to learn about it.