Music Behavioral Data

I figured I’d write up a little response to Fred Wilson’s post on behavioral data. Behavioral data, specifically data collected automatically (as opposed to manually entered by the user), has long been an obsession of mine and was a major reason why we built BlogRollr for Fred.

There are two key components to working with behavioral data:

1. 1. Data collection – Getting your behavioral data into some digital format.
2. 2. Data manipulation – Playing with your data. Analyzing it. Using it to learn more about yourself and others. This is the fun part.

Im not terribly happy with how music applications are addressing either.

Data Collection

Credit where credit is due. Last.fm does a good job of getting my basic music behavioral data online. But it stops there. And they accomplished that like 4 years ago. Not a lot has happened since. It doesn’t get my additional metadata from iTunes. It doesn’t let me vote up songs that I happened to hear on a cd in my car. The concept of a scrobble hasn't evolved into anything more interesting than a play count.

Data Manipulation

I've built up a pretty good chunk of data over the last few years and I just feel like I can’t do anything fun with it. Last.fm pretty much stops at charts. Simple queries such as “What is the list of Animal Collective songs that FictiveCameron has listened to?” are impossible. Their API is weak and does little to foster innovation. It seems like a cautionary tale of what can happen when an interesting startup gets acquired. They built the scrobbler, but have done little with our scrobbling data.

Towards Openness

Id like music apps to pay a bit more attention to the “tweet”, which has emerged as the exemplar open data object. I can take my tweets with me wherever I go. Twitter’s API has enabled a flood of interface innovation / sophistication in a very short period of time. You don’t like Twitter’s interface? Pack up and go to Tweetie. Don’t like that? Check out Tweetdeck. Don’t like that? Go for Twanquility (our product). Why can’t I do this with my music data?

This openness is going to be essential as music heads for the clouds. Ubiquitous wireless connectivity will usher in a new era for music and a movement away from massive file libraries. I want to be able to overlay my music behavioral data on whichever cloud music app I choose but something tells me i'm going to get pwned. Too much data is already wrapped up in closed services like Last.fm and the switching costs are pretty high. No one wants to lose years of data.