All of that's possible, and really, not all that technically challenging... it's just difficult to get community involvement on a large enough scale to make fine questions, like "How does this HAR combination fare against this one", unless they're really dominating the metagame. Futhermore, it's hard to rank data; for instance, if I really wanted a 'meaningful' comparison of head to head matches, I'd limit myself to matches in the cut (where every game is single elim, everything is on the line, and players are already demonstrably high-calibre).
The trouble, of course, is that this cuts our datapoints down to roughly a quarter of what I have in the Annals, so it's very hard, once all the data is there, to answer any specific questions about any given deck type there.
Your best bet would be to pull in data from some source like OCTGN, and I know there was talk of this previously. You could potentially coordinate with tasslehoff to incorporate a 'report match' query when leaving the game. Once the data is collected, even if it's just H/A/R (W/L) vs H/A/R, something as simple as a spreadsheet like the Annals could potentially store the data for public use, and answer 5-10 choice questions that are relevant to the meta. Making one where people can choose there own questions (Does stark N/A fare better against lanni pbtt if it uses meera or crossing)... that requires a little more power than google docs can provide.