It’s hard to predict this in advance, since it’s sensitive to things like voter turnout in non-competitive states. For instance, a blizzard in New England could affect the popular vote without impacting the electoral college vote. So I’ll just tell you how to calculate it.

First, identify the tipping point state. Some guys think it might be Michigan this year. Meaning either candidate can win by winning Michigan and every state more favorable to them than Michigan. Other good guesses are Pennsylvania and maybe Arizona. Then take the difference between a candidate’s margin in Michigan verses their national average.

2020: The tipping point state was Wisconsin. Biden won WI by 0.6% and nationally by 4.5% representing a 3.9% electoral college advantage to the Republican.

2012: The tipping point state was Colorado. Obama won CO by 5.4% and nationally by 3.9% representing a 1.5% electoral college advantage to the Democrat.

So the electoral college doesn’t intrinsically benefit the Republicans, but it probably will this year.

Here’s the list of states and electoral college votes:

https://www.archives.gov/electoral-college/allocation

Just don’t think about the popular vote. It has no bearing on who wins the Presidency in the US.

You can argue that it should but just accept that under the current rules it does not.

As far as your question “If Kamala wins the popular vote, how much does she have to win by to flip the electoral college to her side?” the only answer is “it depends”

It depends because as amazingly stupid as this sounds, one vote for a candidate counts either more or less depending on which state it came from.

Example of California (most people) and Wyoming (least people)

    California:
        Electoral Votes: 55
        Population: 39,500,000
        Weight:  0.00000139

    Wyoming:
        Electoral Votes: 3
        Population: 580,000
        Weight:  0.00000517

A vote in Wyoming (0.00000517) affects the outcome of the electoral college much more than a vote in California (0.00000139).

Another way of looking at it is that one electoral college vote in California represents the will of a little over 718,000 residents, while in Wyoming it represents the will of a little over 193,000 people.

Things get even trickier when you factor in the fact that some states split the EC votes based on popular vote or district, and other states are a winner-take-all (whichever candidate takes the state takes all the EC votes.)

It’s a giant complex mess and it cannot be easily related to the popular vote.

Stop worrying yourself with the popular vote. You gotta get them purple states to WIN.

It doesn’t work that way.

There isn’t a single popular vote. There are 538 votes to be cast, votes distributed to the states based on the most recent census.

Each state then runs an election to determine how their vote is cast. More technically, you’re voting for a specific elector, but given faithless elector laws, it’s defacto a specific candidate.

Most states are all-or-nothing, whoever wins that state gets that state’s full vote count.

To win, it takes 270 electoral votes.

This is why battleground states are so important. Hillary may have gotten more votes nationally, but none of the elections were in fact national. She ignored key battleground states and arrogantly expected to just win them.

Hillary was, and remains, an arrogant fool. Not, it must be said, in the same way Trump is.

Depends fully on what states she wins

This will help you a bit to answer your own question: https://www.270towin.com/ You can play with scenarios there.

how much does she have to win by to flip the electoral college to her side?

This misunderstands how our elections work. If you win a state by one vote (offer not valid in Nebraska and Maine), you get all that state’s electoral votes. If you flip several thousand votes in a few states, Trump wins the 2020 election even though Biden had around 7,000,000 more in the popular vote total. Also, if you’re from the UK, Americans are a little idiosyncratic in their voting. A good deal of people do the equivalent of voting for a Tory even if they wanted Labour in power. It’s called vote splitting, and it’s generally a terrible idea.

The practical answer is 3-4% above to counteract the right-wing effect of the electoral college. Yes, it all matters on what states she wins.

The theoretical answer is that Kamala could get less votes, just like Trump did in 2016, and still “win.” It’s not practical because the swing states are more conservative than the median population of the country as a whole, which means it’s extremely unlikely those swing states will vote for Kamala while Trump gets more votes elsewhere.

The places you need to watch are Arizona, Michigan, Wisconsin, Georgia, Pennsylvania, Nevada and to a lesser extent Ohio, Minnesota and Florida. The 538 polls will give you a sense of where those states are leading, and you can see different maps here. polling is imperfect, and frankly I can’t take the anxiety of watching that data day-to-day.