Failing to find music on the fediverse
Inspired by Martin's 38C3 talk on search engines I spent some time today taking up the idea of using Mastodon's trending links API to see if I could find some interesting music-related links.
I did this by fetching https://www.fediverse.to/data/dataset.js and stripping the javascript to turn it into json and then extracting just the mastodon hosts with a "music" category:
cat dataset.js | jq -r '.[] | select(.cat[] | contains("music")) | select(.sw=="mastodon") | .url' > hosts
I then looped through all 82 hosts to fetch "https://%s/api/v1/trends/links"
Of the 82 hosts, only 15 had any links. And none of the links were music related, most were popular news stories.
It was a nice idea, but it looks like you'd have to actually crawl the instances to collect links if you want anything other than the popular sites.
hacky go code if you want to replicate - https://pb.envs.net/?264dc776fc8d4a69#H2wQEFt8fcGPSvRWmekigHuHqbSb8CABwd69FzKsatsP