Sombody Lovs this Album

[ code, music discovery, API, digital ]

Album recommendations from user generated tags/genres with typos from last.fm.

Sombody Lovs this Album is a recommendation generator that pulls from a collection of last.fm tags that are misspelled words of affection. It acts in opposition to streaming platforms that recommend based on that they know about your listening listory and outside of the landscape of sponsored playlist placements.

BACKGROUND

On last.fm, users can connect their streaming service of choice and the platform will track what they listen to and aggregate this data into a user profile. There’s a social network component as well, where you can see what your friends listen to.

Users can add “tags” to what they listen to (like a hashtag) to organize their collections (like “favorite album” or “rock”). Because these tags are user-generated (meaning they don’t come from a dropdown list, but rather whatever the user types), people make typos. And sometimes, a bunch of people make the same typo. They forget the “e” in “awesome,” they forget the “u” in beautiuful,” etc. This project pulls a random album with that tag. 

MATERIALS & PROCESS

p5.js, last.fm, paper. I accessed the tags through last.fm’s API and created the sketch in p5.js. Tags include words like "lov", “awsome", and "gorgeus".

EXPERIENCE

When the user reaches the end screen with their recommendation, they are instructed to write it down on the paper provided, instead of taking a photo or writing it on their phone. The end result is a tangible object to refer to later. When I shared this at ITP Camp, I provided index cards to write down the artist, album, and tag name.