Have you done any fine-tuning or prompt-customization for the R-specific work? I've found the models worse on R when compared to Python, especially for more complex tasks. This looks cool, thanks for sharing!
Nothing R specific. In my experience, Claude is pretty good about using tidyverse for everything. What was is flopping on for you? Our thought on not fine tuning models is that whatever comes out in 6 months is just going to be better than whatever we fine tuned.
Admittedly this was probably ~2 years ago, previous gig used a lot more R and the earlier models were (in my memory) worse in R than in Python. But makes sense that this would've come a long way. Love this!