I don't think that this is an accurate method for of determining this, but Nim does fairly well in it. Obviously Nim lacks popularity, but Nim is ahead in expressing ideas in less lines of code. See this comparison.
I'm a big fan of programming language comparisons, and I'm happy to see one trying to measure terseness.
I personally would measure terseness by first stripping out comments and empty lines, normalizing all identifier names to same length, and then counting characters weighed by typing effort (ex. 1.0 for lowercase, 1.2 for numbers, 1.4 for uppercase, 1.6 for punctuation, 3.5 for unicode, 5.0 for line break). This would adequately punish $perl, P_H_P, and especially languages that make you type "endnendnend". 😏
But it doesn't make sense to me to graph two totally disconnected things, popularity and terseness, as the X and Y axis of the same graph.
It would make more sense to graph two things where there's a trade-off between them, like:
Then you could cluster languages into categories: highest-performance with high-effort, lowest-effort with low-performance, unrewarded verbosity, and the contest to find the language that provides highest maximization of both (which is where Nim would excel).
It would also be interesting to compare different popularity rankings against each-other, like TIOBE (which approximates Web content quantity) vs PyPL (Google Trends) vs Module Count Growth vs specific community counts (GitHub Pull Reqs / Pushes / Stars / Issues, Reddit headcount, IRC, StackOverflow, etc). If some of the latter measure early adopter enthusiasm, then that can be used to predict future mainstream popularity growth.
It would have been great if Rosetta Code (or a blatant scrape thereof) would let people rate the aesthetic qualities of a correct implementation. No measure is perfect, but that would be more valuable than calculating code "terseness".
(I hate to endorse a proprietary evil powerhouse like GitHub, but having users log in with their accounts there, would even allow weighing or discarding votes based on their stars. Sites like IMDB do something similar to weigh user reviews.)