There is a fundamental flaw in this analysis. It assumes that fumbles are randomly distributed.
When there are skill factors on both sides (causing and preventing), the base distribution is certainly not normal, for any team.
The data compares all teams on fumbles. Unless you believe that the Pats are super human compared to the whole rest of the NFL for being nearly off the chart, then the obvious conclusion considering they finally got caught with deflategate is that Pats were cheating.
Also look at the fact that the Pats play at home in an open stadium as mentioned in this article and now you are talking about the Pats being super super human.
This data analysis is actually very good.