Yeah, I definitely think it's measuring something. It's data, so of course it's accurate in the sense that however they happened to measure type it seems to predict that those who score ENTJ will make way more money than those who score as INFP. The bit that's questionable is whether the way they measured type was really measuring type in a useful way. Like if one of their questions meant to evaluate presence of Te was "Do you feel you have achieved a lot in the workplace?" then of course that's going to correlate with financial success. Is that a good question for measuring Te though? It might be. I'm not saying that it isn't, but just that there may be any number of questions which are unintentionally biased in such a way as to inevitably correlate X type with Y financial outcome.
Sorry that was kind of long-winded, but I do agree with you ultimately that the graph is probably at least sort of right about a few of these. I can say anecdotally that, as predicted, I am certainly not making that much money