Fantastic news! So to add the llm.txt to my site, which is built, rendered and deployed via GitHub Actions, I just switch to the new version in the workflow or do I need to explicitly call build_llm_docs() after build_site_github_pages()?

And a Makefile to orchestrate the execution order of different reports and their potential dependencies ✌️

You can still publish to shinyapps.io from Positron using the rsconnect package

Another challenge is being obsessive about function names and consistency since in R we typically load everything into the same namespace. In Python, I’d probably take a different approach, since it’s more common to organize things into (sub-)modules.

Great idea for a post! The tricky part is that you often figure things out along the way (e.g. column / parameter names), which then forces changes in other packages (i.e. breaking changes / soft deprecations). A meta-package that loads the rest is easy to set up thanks to the tidyverse 🙏

Link to package docs: tidy-intelligence.github.io/r-econdatave...

Wasn’t it some Excel mess? Like hitting row limits and wrong references. Guess the first step would be a code-first approach 🫠

The first vignette offers a detailed look at what’s possible with {datacommons} in #rstats: tidy-intelligence.github.io/r-datacommon...

Working hard at more R packages 😅

What’s the problem with webR in Safari? How about Chrome?