Since I run https://www.music-map.com this naturally caught my eye.
Unfortunately, there is no word on the methodology except that they did it with "the help of YouTube’s geocoded streaming data". Does anybody know what that is?
Unfortunately, there is no word on the methodology except that they did it with "the help of YouTube’s geocoded streaming data". Does anybody know what that is?
Also, I'm guessing that the nyt got some sort of special dataset from YouTube. Geo located streams are priveleged to the content owner, per the YouTube API.
Frustrated with music recommendation services I built a ruby script this past weekend to scrape my favorite artist's last.fm pages and then iterate on each follower, compiling a list of their likes to sort of crowd-source music recommendations for myself.
This too resulted in failure, it seems that the problem is me, I'm far too picky.
I'd love to get your insight on any alternative strategies for finding new artists if your likes don't necessarily fit squarely inside genres.
Go to https://artists.youtube.com/insights/location and click on "How is this calculated?".
Is the scale relative? Is there any indicator how scaled up or down the relationships are?
Each artist has a desired distance to each other artist and is trying to get to a position that satisfies all these desired distances as good as possible.