After my perspective on electoral vote and popular vote margins--in which I argued that the popular vote is meaningless--I thought about how Electoral College and popular vote margins related to one another. I took a stab at a visualization by creating a slopegraph.
This was much more challenging than I thought. And perhaps it's more deceptive than informative. But why not give it a shot and let the critiques come....
I wanted to show the relationship between electoral votes and popular votes. I started by taking the raw popular vote totals of each candidate--this could have been as a percentage of electoral vote, but 1968 really screwed things up and messed with the visualization if I were using the raw electoral vote totals as the left data point, so I took the slightly less perfect version of the raw vote totals. I started from 1944, which had just 531 electoral votes, in comparison to today's 538, and some other deviations along the way.
Then I opted for the percentage of the two-party popular vote margin, which was also imperfect as a kind of comparison--it might lead to significant fluctuations if there is a particularly significant third-party candidate who draws votes disproportionately from one candidate.
In order to do the slopegraph on two different Y axes, I opted to calculate Z-scores for each side. That offered the relative performance between electoral votes and between popular votes, and it offered some comparable scale between the two from 1944 to 2016.
You can see a couple of significant differences between the electoral vote "landslides" of 1972 (Nixon winning 520 electoral votes, dark green) and 1980 (Reagan winning 489 electoral votes, light green). In '72, Nixon snagged a whopping 61.8% of the two-party popular vote. But in '80, Reagan secured just 55.3% of the two-party popular vote.
There's not much of a rhyme or reason between the performance in the Electoral College and the popular vote--except that we might notice particularly low-performing popular vote winners: Bush in 2000 (271 electoral votes, blue) had the razor-thin electoral advantage; somewhat healthier were Trump in 2016 (306 electoral votes, pending December 19, red) and Kennedy in 1960 (303 electoral votes, orange).
In any case, perhaps after all the flaws I've identified and the meaningless of the popular vote, anyway, such a slopegraph is of less than even marginal value. But here it is, if you find it of interest.