I wonder if the same happens to you. Finding anything relating to corpora research on the Internet is virtually impossible to me.
If you try to find an estimate of the data (mime) types being transferred over the Internet (text: html, pdf; video: mp4 ...) all you get is ads of companies trying to get your money, sell you "solutions" ...
I think those kinds of statistics at least, approximate figures, should be somewhere. If not the data transfers per se, the links on pages.
Do you have an idea about where to find such data?
lbrtchx