For a control group, I picked the top 25 adjectives in the English language, all of which are relatively non-sensational words like "good" and "new" and "different". For our purposes here, this should work just fine.
As in Ignorant Jane's original (and hilarious) post, I did a Google search for "site:io9.com [adjective]", then followed it up with an unconditional Google search for just "[adjective]", recording the number of results pages for each.
On average, io9 used each of these "control" adjectives 0.000047 times as often as the entire Internet. I used this average as a baseline 1:1 ratio, and from that, produced the following graph:
(Click to embiggen.)
Just in case you're confused, the "1.00" means io9 is using the adjective just as often as the Internet as a whole. Io9 uses words like "bad" and "few" more often than the Internet (1.89x and 1.91x more often, respectively) and uses words like "public" and "large" less often than the Internet as a whole (0.36x and 0.42x as often).
But overall, io9 uses these "control" adjectives roughly as often as the Internet uses them . . . nearly twice as often in a few cases, but constrained within the same order of magnitude.
I then found a thesaurus reference for " synonyms of awesome". There were 24 highlighted synonyms (including "awesome" itself) so I performed the same searches using all 24.
Here's the "awesome" graph:
The weighted average for all synonyms of awesome is "2.75" . . . that is, io9.com uses "awesome" or one of its synonyms at least once in an article at a rate 2.75 times higher than the rest of the Internet uses those words across all webpages (when compared with how frequently io9 uses the control group adjectives).
"Awesome" itself is used 3.5 times more frequently. "Stunning" is really not that bad . . . only 1.7 times more frequently!
Take this for what it's worth (which is nothing). I took one statistics class in college and a real statistician would probably vomit looking at my work here.
Oh, and here's images of the raw "awesome" data:
And the raw "control" data: