Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Spotify’s Discover Weekly: How machine learning finds new music

hackernoon.com

Post reply on HN
Print view View original

131–140 of 275 posts

Re: Spotify’s Discover Weekly: How machine learning finds new music

#131

Post by freetonik »

kolp wrote: post #16

Spotify's discovery engine playlists (Discover Weekly and Release Radar) just don't work for me. I now ignore both playlists and I suspect that the lists are influenced by payola. I've tried training the algo by following artists and saving albums in the style that I would like, but these playlists keep peddling stuff that is way off the mark. Interestingly, the daily mix playlists have responded to this training, bu…

> Once a song is in the archive, it won't (or shouldn't) appear in either of the weekly so-called discovery playlists.

They still do, sometimes, at least for me. But more importantly, adding songs to playlists sends the positive signal to Spotify’s recommendation engine, as far as I know. So, saving all your discover weekly tracks into archive playlists will encourage Spotify to give you similar music in the future.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#132

Post by landtuna »

steve-benjamins wrote: post #89

As an artist, Discover Weekly has been the best thing to ever happen to me. Every Monday I get a big infusion of listeners (around 5,000)— many of who stick around and check out my other music :) Prior to that, the best press I could get was the tedious process of cold-emailing bloggers (a practice which is now dying off).

Haha - read your comment, looked at your username, opened Spotify, and "Purification Ritual" is the first song in my Discover Weekly.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#134

Post by eru »

probably_wrong wrote: post #104
Fabled wrote: post #57

In case you didn't know, Spotify categorizes music into genres behind the scenes. You can use this site to find out what genres your favorite artists are categorized as: http://everynoise.com/engenremap.html You can use this information to then check out Spotify's auto-generated playlists for each genre. They have at least three types for each one: "The Sound of ", containing definitive representation of the genre, "…

> You can use this site to find out what genres your favorite artists are categorized as That's my problem right there: my favorite song right now is a musician's piano cover of one of his own songs. His music is usually electronic, which I don't like, but I love this one song. So Spotify will recommend me electronic music from other artists, which of course does not fit my song. Repeat for every author. I liked one…

Pandora works song-by-song (and is not bound by genres).

Re: Spotify’s Discover Weekly: How machine learning finds new music

#135

Post by Pigo »

ikornaselur wrote: post #90
arkadiytehgraet wrote: post #80

If there is at least one song that I like in Discover Weekly, then I consider it to be a good mix. However, this happens once a month at best. After reading the article and all three methods for recommendations, I cannot understand how on Earth it comes up with its suggestions. Do I feed it bad data? Does it not take into account the relative "weight" of a particular track, i.e. my preference to listen to it multiple…

Yeah, I was surprised when I read the start of the article. On Mondays, I will start the day by playing my discover weekly playlist and I'll usually have it running throughout the day, saving 1-4 songs from it to my library. Sometimes none. I just checked and this week I didn't even save a single song for example. I guess it doesn't help that I do listen to really different types of music depending on what I'm doing.…

I don't really use Spotify. But this is why I like Google Play Music, because the playlist suggestions are categorized by what you're doing (rockout at the gym, christmas, working-chillout, working-fun).

But honestly you can't beat just going to youtube for a live stream if you want some chillout-ambient music while you're programming. There's so many talented dj's out there culling new music.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#136

Post by emodendroket »

arkadiytehgraet wrote: post #80

If there is at least one song that I like in Discover Weekly, then I consider it to be a good mix. However, this happens once a month at best. After reading the article and all three methods for recommendations, I cannot understand how on Earth it comes up with its suggestions. Do I feed it bad data? Does it not take into account the relative "weight" of a particular track, i.e. my preference to listen to it multiple…

Mine is pretty good except for it thinks I want to listen to like ten Scandinavian language songs in a row.

Although now that I think about it, one thing that seems to lead to wacky recommendations is sparse data. If I check out Japanese enka albums the "similar to this album" information has stuff like Justin Bieber, which mostly tells me they have so little data on what kind of listeners enjoy the music that it's pure noise.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#137

Post by jefffan241 »

Pigo wrote: post #135
ikornaselur wrote: post #90

Earlier quoted context omitted.

Yeah, I was surprised when I read the start of the article. On Mondays, I will start the day by playing my discover weekly playlist and I'll usually have it running throughout the day, saving 1-4 songs from it to my library. Sometimes none. I just checked and this week I didn't even save a single song for example. I guess it doesn't help that I do listen to really different types of music depending on what I'm doing.…

I don't really use Spotify. But this is why I like Google Play Music, because the playlist suggestions are categorized by what you're doing (rockout at the gym, christmas, working-chillout, working-fun). But honestly you can't beat just going to youtube for a live stream if you want some chillout-ambient music while you're programming. There's so many talented dj's out there culling new music.

Spotify has playlists for the same things you listed (minus maybe christmas, but I'm sure they do). I've seen the others though listed.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#139

Post by emodendroket »

KozmoNau7 wrote: post #124
mixedbit wrote: post #121

Earlier quoted context omitted.

My friend likes and uses the Discover Weekly and to get good recommendations he uses different Spotify account when he plays something for his kids or plays some generic background music at home.

He can use the "private session" option for that instead.

I wish that option were more easily accessible instead of being hidden away behind multiple menus.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#140

Post by markussss »

The biggest problem with the Discover Weekly is its inability to understand _why_ you're listening to a specific subset of music. It might not be that your taste in music suddenly changed, or that you discovered a new genre that you're incredibly interested in, even though your most listened to genres or songs changed for a few weeks, or your listening patterns changed for a few weeks.

A few examples:

I'm Norwegian, and listen to quite a lot of Norwegian music in Norwegian. Norwegian music is also European music, Scandinavian music, Nordic music, etc., and as a result I get music from Germany, France, Finland, Iceland, Sweden, Denmark, among other countries and languages. However, the reason that I enjoy listening to Norwegian music is because I speak and understand the nuances in the language perfectly, while this is not the case with any of the other languages.

In Norway, we have «russefeiring» (russ celebration) (https://en.wikipedia.org/wiki/Russefeiring) from approximately mid-April to mid-May. In the recent years, many groups of «russ» have been making/ordering songs to represent them throughout their celebration, and in that period, I listened to some of those songs because it was fun during that period (around April-May), but it's not interesting to listen to outside of that time period at all, pretty much like Christmas songs. Now it's the second week of October, and my Discover Weekly list still contains a lot of songs (12 of 30 songs) created specifically for the «russefeiring». Imagine getting half your Discover Weekly filled with Christmas songs in May, because that's pretty much my experience with this.

I don't know what they have to do to make Discover Weekly not give me shitty suggestions, but right now it keeps giving me suggestions that are outdated and uninteresting and I have no good way to give that feedback.

I'm really saddened by this, because for the first half year of Discover Weekly, the playlists were so good that I stored them in separate playlists to be able to go back and listen more to them.

Post reply on HN