petebachant.me

Evolving iPod playlists

September 08, 2010 | 1 Minute Read

I’ve owned my 30 GB iPod for over 5 years now. I leave it in the car and really only use it there. It contains 4168 songs and I find myself constantly skipping through when it’s on shuffle mode. It’s pretty clear I don’t like all the music on there but I’m too lazy to find and delete the unappealing songs or create manual playlists from my favorites. I just recently discovered that it’s possible to make smart playlists in itunes using number of skips as a parameter. This allowed me to fulfill a dream of mine, a playlist that evolves. The idea is like artificial selection except there’s no reproduction.

Songs are split into tiers based on number of skips. Tier 0 contains songs that have never been skipped. Tier 1 contains songs which have been skipped 1 or less times, and so on and so forth. Now as the iPod is used, the tier 0 playlist becomes closer and closer to the listener’s “favorite” songs. Of course, newer songs which haven’t had a chance to play during shuffle will be included, but if they’re not “fit” they’ll certainly be gone soon.

After I made up to tier 3 I wondered how many tiers it would take until all songs were included. It took a whopping eleven! I plotted percent of total songs versus tier in the figure below. It’s an interesting curve. Note that roughly 12% of the songs have never been skipped.

Evolving playlist

I’m pretty excited that my iPod will now get “better” over time. Good move by Apple including this data logging feature!