has anyone with access to the data considered doing predictive/explanatory analyses rather than pure descriptive?
I've seen it mentioned that the data is being released with an eye toward developing player performance metrics (ie things that will be useable by clubs above and beyond pure description)
I'm a research scientist (in NZ) with some football knowledge (above average but not spectacularly knowledgeable) and was considering starting a project looking at developing non-observable underlying performance measures, - in lay-speak - taking a bunch of variables that are all related and seeing if there are any non-observable (latent) factors underpinning performance across the variables/time.
I was considering developing metrics via factor analysis or PCA (principle components analysis) but don't have the breadth of knowledge in football to intuitively know manifest (observable) groupings of skills, abilities, attributes etc (where they would exist as underlying assumptions or be used to theoretically underpin hypothesis)
This is purely speculative at the mo, but as I work at a high profile sports science uni here in NZ, I would be interested in collaborating with (or just info sharing/learning from) knowledgeable football types with an interest in analysis with the aim of developing publications from it (preferably theoretical and context free so as not to violate the terms and conditions under which the data were released)
anyone on here been thinking something similar or have ideas around this sort of analysis and want to go further with it?
(and no i wasn't at York away)