There’s a row in PA listing Biden with 104.3% of the vote in a block with 1285 votes (the differential increased by more than the total number of votes). Data error?
I also see 3 blocks in PA of 25k+ votes with 90%+ Biden (one is 97%) - can someone better at statistics tell me how likely that is due to chance given the other values (assuming the data is reporting accurately)?
Expected. Last time round, Clinton got 82% in Philadelphia. And mail-in voting was majority Democratic in most places this year.
You'll see similar, on the face of it absurd numbers in the other direction in very red areas for the in-person vote, where the few Democrats available mostly voted by mail, so virtually everyone voting in person voted for Trump.
Philadelphia normally goes for the Democratic candidate by around 80%, and mail-in voters this year are disproportionately Democratic, so it is not unusual that some blocks of main-in ballots in Philadelphia county would be 90% Biden. Note that the total vote (mail-in + election day) in Philadelphia county for Biden is exactly 80%, so there doesn't seem to be anything anomalous here.