Mmm, very good points. I used R for quite a few high-level statistical analyses, I found it handy for munging data and .rmds are nice for organizing a systematic set of chunks to make the code logical. Sometimes I ran into speed issues, which is when I would reach for data.table, which is quite a huge speed boost for some of R's slowest issues.

I had to learn the basics of STATA for a modeling project, because the other researcher knew STATA (these network effects of collaborators knowing different things). Turns out fixed-effects multivariate logistic regression modeling is WAY faster in STATA than R, to the point where our infra just couldn't complete the R code in any reasonable timeframe so STATA was just the better pick.

My main gripe is especially in healthcare, so many workers are trained in something in their 20's, then use that for decades. There are thousands out there still using SAS every day, for instance, because it's what they and their collaborators know. I suppose Julia and Python will grow in healthcare over time, SAS should in theory go down in usage...who knows!