Hey,

The current project I'm working on requires an amazing amount of detail on what specific users are doing (we're talking proper research levels of data). Ideally we want raw, unfiltered data i.e. in no way preprocessed by a 3rd party.

I've had a quick look at Flurry (and used it for a project or two) and unfortunately it doesn't seem provide data on specific users, but rather only averages. So it doesn't seem to give me a per user breakdown of events for instance.

So to give you an idea of what I will probably need, per user:

[User id]:
o Total Play time
o Num Sessions
For Num Sessions:
o Session Length
o Successes
o Failures
etc.

The people I'm working for will want to be able to draw complicated graphs with standard deviations and what not.

I was wondering if anyone had already looked at doing anything like this?

It's going to be pretty easy to accumulate that kind of data as the player is playing and just storing it in an sqlite database or json flatfile, the issue is going to be getting at the data.

At the moment I'm thinking that this will involve me writing a basic Python webserver and using the Async HTTP to post data to it at regular intervals.

Any thoughts?