Why would you omit 2 seasons worth of data though? I mean, respectfully, I'm not sure any data set has much value if we selectively choose what we do or don't include. That's kind of the point of data. It will give you conclusions that may ultimately challenge our pre-conceptions.
If you are including it, Southampton V Man City is the most watched PL game of all time. I think we all agree that Covid, through the BBC airing of some games and the very nature of the % increased TV eyes distorts the data.
Also, not to mention the unprecedented hiatus that lead to incredible viewing figures on return, the derby being the highest I think.
I have no doubt the bundesliga saw huge viewing in the UK relative to normal when they kicked back off post lockdown 1 - but it doesnt make that the norm.