I'm new to COLLADA and apologies if this question is obvious ...

Given a set of data - say aircraft tracks - what is the best way to transform it into COLLADA so it can be consumed? I'm considering using an ESB to do the transform (since it is already in the infrastructure) and either XSLT or XQuery.\

The idea is that data is coming in non-stop, the ESB sees updates and automagically transforms into COLLADA so it can be visualized in real-time.

Has anyone done something similar to this before?