last.fm

Last.fm lets you see your 'neighbours,' people whose music tastes are similar to yours. I've been fascinated by this list for a long time, and have taken to reading through the favored artists of each 'neighbour' to see if they already knew of artists or groups that I should be listening to.

It's actually worked. I first encountered Snow Patrol through a neighbour, and Adam's recommendation clinched it. (Though that's perhaps a misleading statement, because Adam and I trade music recommendations so often that he's been a regular on my neighbours list for some time now.)

It hit me, though: for the most part, neighbours were always slice-of-life snapshots of similarity. Single notes, if you will. If you picture a person's music taste as a chord of notes, then can we all agree to beat this metaphor into the ground and say that most of the people on the neighbours list had musical tastes that matched only a note or two in my chord.

In the spirit of Valentine's Day, and by that I mean "this doesn't have anything to do with Valentine's Day," I present something I love:

Open data formats.

Most of you know I've been logging my listening habits over at last.fm for several years now. Some time ago, a nifty gentleman wrote a set of scripts (available at lastgraph.aeracode.org) to generate a visual representation of a single user's listening habits over time.

2005

2006

2007

I think I have earlier data saved on files at home, but I'll have to check. Clicking on the files will take you to the larger versions archived on flickr.