Each entry in the database contains a language and a number of pages. I sorted all the entries by language and took the average number of pages for each of them. But it also display a major weakness, each language don’t have the same number of entries, some have thousands, others less than a hundred. I should have “normalized” the number of entry for each language and exclude languages which don’t have enough entries.
I live in France and I can confirm, this is how we drive roundabouts (but you have to watch out for the people demonstrating (or not🙃)).