If you want a somewhat scientific approach to predicting how states are going rather then getting all excited when a giant portion of a state come in that clearly don't represent the state as a whole.

I recommend
https://www.nytimes.com/interactive/2020/11/03/us/elections/results-president.html?action=click&pgtype=Article&state=default&module=styln-elections-2020&region=TOP_BANNER&context=storyline_menu_recirc

They wont show a forecast till a decent represented of state have reported (that could mean just 1% through of different locations). The percent will shift greatly early on because any one place someone under or over performing early going to shift it greatly. For example at time of this post there a forecast for Georgia but not for Ohio even through more of Ohio report. That because it not yet a good representation of the state.

Site does require a free account.

Eh, how is Biden leading in Kansas with 32% counted by 54% to 44%? Kansas?

Also I can't believe Texas, but that is close and less than half counted so far.

Remember every state has areas that are blue/red. It doesn't mean anything if you see a lead. If you want to see if things are different, look into county data, not state data.

It also helps to look at which counties are winning. While not directly like the EC, you get an idea of how good performance is by seeing if they are taking counties they otherwise missed before. 

For one, Biden's Texas lead has started to erode as rural ballots start arriving. 

Florida Hispanic vote is currently 53-45 for Biden.
It was 62-34 in 2016.

Kansas I don't understand either (can't check the counties, otherwise I would have some insight here, but it's not updated yet)

Texas however is understandable. Biden is polling ahead of Clinton, who was already pretty close in Texas, by about 2% in urban areas and 8% in rural ones. That's a huge uptick for the democrats and if this trend continues for the other counties then Trump will lose the state by over 3 points.

Hispanic voters per early 2020 exit polls:

Florida:
2016: Clinton +27
2020: Biden +8

Georgia:
2016: Clinton +40
2020: Biden +25

Ohio:
2016: Clinton +41
2020: Biden +24

CNN

— Political Polls (@PpollingNumbers) November 4, 2020

If you search in Google 'US election <state>', you get actually a map with counties. For Kansas I see, two counties in blue, which were light red in 2016. Clinton won 2016 only two counties, both of which aren't counted yet. So Biden apparently does better. But... the real dark red counties are all still missing. So Kansas probably will switch around, even if a third is already counted.

Alright, so this is what I got.

There are only two blue counties right now:

-Johnson: With 81% of the vote recorded, Biden is up 17 points. Trump won this county by 3 in 2016. Biden has 20k more votes than Clinton already, but expect this margin to narrow with time.

-Riley: With 63% of the vote recorded, Biden is up by 25 points. Trump won this county by 3 in 2016. Biden has roughly the same amount of votes as Clinton in 2016, but expect this margin to narrow with time.

This is a pretty big shift, but don't expect Kansas to turn. The issue is that the more red counties have very few votes counted right now, and mail in votes are reported first, so we WILL see a red shift as in person votes are tallied. 

So a pipe burst in Georgia and they have stopped the counting. lol