Been noodling around on the table and something that I think is far more interesting is when you start looking across races rather than between them. For example, if you ignore the tops and bottoms of some of the races, they seem to be spread fairly evenly (more or less) across the classes. Humans (ignoring fighters which is a huge bump at the top) are more or less even across the other classes - as expected. Tieflings are another good example, ignore Tiefling Warlock at the top and the spread across the other classes isn't all that huge.
This might be a good measure of how successful a race is in the game. Goliath, for example, is basically Barbarian, then everything else. It's very much a one trick pony. I mean, about 50% (ish) of goliath characters are barbarians. Tag on fighters, and you've got a bit less than 2/3rd of all characters. Probably not a great candidate for elevation into the PHB. Too one dimensional.
Conversely, Dragonborn have a really nice, smooth curve. Yeah, Dragonborn Paladin is at the top, but, if we remove top and bottom, the spread of classes is 75% of Dragonborn characters. That's a race that has a very broad appeal. And, this is a really good reason to keep half-elves as a separate entry in the game. Again, a race that has such a broad appeal, drop of bard from the top and barbarian from the bottom and that's barely even a curve. That's a pretty flat line across all the classes.
And, to be perfectly fair, halfing, once you drop off rogue from the top and paladin at the bottom, is actually pretty evenly spread. And the rogue bump at the top isn't all that huge. 1/3 of halflings are rogues (give or take) and the other 2/3rds are something else. Which, to me means that the race is actually pretty well rounded. Certainly better than goliaths in that respect.
It's just another way of looking at the data.