TuneGlue

TuneGlue is yet another service that aims to remix usage data from the likes of Last.fm to reveal patterns and relationships between the acts we like. I can't say that it provided me with any startling revelations when I fed it the names of some of my favourite acts, but the graphical user interface used to manipulate acts and reveal more in-depth information is rather impressive.

If the same interface was used with an individual's profile, rather than with aggregate data from all users of the services it pulls data from, it might be more illuminating. When I asked it about Pulp, the acts it revealed were New Order, Blur, The Tears, Suede, Blur, The Stone Roses and Manic Street Preachers. I'm sure that general sales patterns do point to those relationships, in that lots of people in the mid-late 1990s would have been buying albums by more than one of those acts, but I don't think there was any particular musical kinship between, say, Pulp, Manic Street Preachers and New Order. I'd expect a service that was trying to identify artists similar to Pulp to mention The Kinks and Scott Walker.

Similarly, when I asked about Kate Bush I was told that the first wave of related artists consists of Joni Mitchell, Tori Amos, Suzanne Vega, P J Harvey, Peter Gabriel and the Eurythmics. Contemporaries, certainly, and perhaps with a slightly greater degree of musical kinship that the previous sample, but for the most part still not obviously artists I'd like if Kate Bush was my thing.¹

Still and all, it's early days yet. As they gather more data and figure out better ways to link it together², I'd imagine that their nifty user interface will come into its own.

[Via Very Short List]

1. Which, if you'll pardon the expression, she most certainly is. ↩
2. In particular, differentiating between usage data from Last.fm and sales data from Amazon. ↩