Thanks for running the data, I love this kind of stuff.
Something I might run later; The top level BBQ (i.e. total number of banners among the top 24 teams). Would that matter more than the average number of banners across all teams?
I know you said there may be little errors... but the regionals I checked had a lot. Minnesota 10,000 Lakes has 2470 listed as earning two blue banners when they've only ever won one award (though if that's a prediction for 2010 I'd say it's a pretty good bet). 1714 has 3, not 2; 2194 has 2, 2826 has 1, 2177 has none (two regional finalists though). It appears you're counting Regional Finalist, though even including that it still doesn't explain 2194 / 2470 / 2826. (nothing against these teams, I'd prefer to see them all with more blue banners)
Perhaps your script is grabbing the wrong awards? I dunno. For the record I used the FIRST event pages to get all of the data I cited above.
As for my opinion on the hardest regionals: Midwest.