Would any of you use something like this? I know I would

I'm thinking about building a service much like Google Music (cloud backup of an audio library), but with a lot of great features to help DJs find the perfect track. You would upload your library (synced automatically) and it could be analyzed for pitch, tempo, etc, as well as metatags. You could add a metatag like "grows slowly" or "heavy drop", etc. Metatags could even be aggregated from other people, and ID3 tags could be automatically corrected if you happen to not have them tagged correctly. Then you could query the system and say something like "I want to mix out of this part of this track and I need something great to mix it into" and it would show you the best candidates and why they were chosen. What do you guys think? I'd really like to hear your thoughts. I'm thinking about building this, looking to see if there's interest in the community.

Seems like a tough one to pull off. Not sure what your profession is, but the technical matter on its own (interface, algorithms, servers) is a pretty tough one to crack. In addition, considere how difficult it was -even for Google - to launch its Music service (considering all the copyright issues).

Would be a pretty cool idea, but then again, choosing tracks is kind of what defines you as a DJ.

You'd never get the cloud aspect off the ground due to to intellectual property issues.
A database of tunes and tags is entirely feasible I guess, though you'd need a critical mass of people to get it to work. The closest thing to it currently is probably 'Lazy DJ', which queries mixes put on mixcloud with a certain song and comes up with a list of tunes people played next in a set.

I don't see why there would be any problems with intellectual property issues. There wouldn't be any ability to share files across the service. If storing people's illegitimately acquired mp3's is illegal then Dropbox is going to have a major problem.

I definitely have the technical background required. I'm a web developer by trade and have spent the last year or so studying musical recognition systems.

That Lazy DJ is definitely an interesting way to go about it! I'll have to check that out.

edit: also, this won't need a large user base for it to be useful for all of the users. There's barely anything social about this. Think about it like Google Music + Mixed in Key + Pandora.