Grammars of Data: lessons from ~20 years of the tidyverse
summary
Twenty years of tidyverse work emphasizes concrete abstractions, tooling, and community practices: tidy data and the grammar of graphics remain useful; the team builds ggSQL: SQL-driven grammar of graphics to let SQL users produce ggplot-style plots; they create a data-dict.yaml CLI in Rust for machine- and human-readable data dictionaries; unit testing and pkgdev tools (testthat, roxygen2, devtools, pkgdown) power package quality; AI tooling like Claude accelerates maintenance—closed 100–150 issues in two weeks with Claude—while Stack Overflow usage collapses and community models are in flux.
AI-written summary · about
chapters (10)
- 0:00Introduction to talk
- 1:05What is the tidyverse
- 2:52Why abstractions matter
- 7:29AI and grammar of graphics
- 9:59Importance of tools
- 15:42New data dictionary tool
- 18:29Role of people in community
- 21:01Future of community platforms
- 22:06Summary of key points
- 23:51Closing thoughts on AI
related talks
51mProgramming in the Tidyverse
18mInteractivity and Programming in the Tidyverse
46mElegant R: Tips from a Software Developer
21mLearning and using the tidyverse for historical research
33mMaintaining the house the tidyverse built
05mImplications of R Syntax in Intro Stats
21mOur colour of magic The open sourcery of fantastic R packages
05mWhy tidy data matters