It's quite dishonest to say that this is a mass survey and draw comparisons between regions when n=600 per champion. This data is worthless, it's just random noise (at least quite close to it).
Only 600 people got a chance to rate how frustrating X champion is to play against per region, and then you compare.
Run the same experiment again in NA and you would get absurdly different result. Something that might happen is that the 24th most frustrating champion to play against might become the 78th most frustrating champion to play against when you run the experiment again (and hence ask different people).
Another thing to note that you're surveying completely random people (or I guess it's not completely random, its from people that played the game in october 2019). Sure, that can be good but it can also be bad. It can be bad in the sense that you randomly survey a bunch of people that barely play the game, you survey a bunch of aram only players (who's opinion doesn't matter). When n=600, it might be the case that a lot of people that were asked about champion X have barely even played against it (since you're asking players that don't necessarily play the game a lot).
You would never take a champion winrate with 600 sample size seriously, why are you taking this ranking "how frustrating is champion X from 1 to 5" seriosly? with just n=600, the average for champion X might be 3.7, but the second time you run the experiment it's 2.7 (which is a massive difference). It's quite volatile with just n=600 per champion, just like champion winrates are.