So much to talk about!
Well, RStudio Conference 2022 has come and gone, and I thought I’d just write a couple blog posts of my experience attending it.
Before I start, I just want to mention the big news: RStudio (the company) is becoming Posit. This new name reflects the fact that a lot of their new products, including Quarto, Shiny for Python, and Tidymodels/Vetiver are cross-language.
Just to be 100% crystal clear: the products, such as RStudio, are not undergoing a name change. Posit is the new name of the company, not necessarily its products.
Day 1: Tidy Tools: Package Building, Unit Testing and Documentation
I was a TA for the Building Tidy Tools workshop, which is a great intermediate course that helped fill in some of the development gaps that we all have. I learned a ton from the workshop, and the other TAs: Beatriz Milz, Haley Jeppsen, Kailey Mulligan, and Elliot Murphy.
Ian Lyttle and Emma Rand taught the workshop1, and did a wonderful job breaking up the material into eas(ier) to understand sections, starting with package development, documentation including {pkgdown} sites and unit testing. I especially appreciated that the entire workshop was structured around building a package called {ussie}2 that let a user work with soccerdata from packages such as {engsoccerdata}.
Day 1 TL;DR: Use the {usethis} package for building your package frameworks and making your unit tests. That and {devtools} are your guides into the wilderness of developing packages.
Day 2
Day 2 started with a more philosophical bent; that of function design. Specifically, the difference between pure functions (that only process what you give them), and functions with side effects (which alter the outside universe3). One of the key insights is that it is way easier to write unit tests for pure functions, and it is harder to write tests for functions with side effects.
We talked more about testing functions with side effects using snapshot testing.
Ian is way better at grokking the tidyeval documentation than me, so it was a pleasure watching him teach the two main uses of tidyeval: 1) data-masking and 2) tidyselect. Specifically, he talked about different ways to implement
I think there is a special bond that’s formed when you teach together, and it was great to work with the entire team.
Day 3: Really Big Announcements
Day 3: Shiny, Past and Present
Day 3: Awesome sessions I attended
Day 4: Quarto is incredible
Day 4: Other awesome sessions
Footnotes
Citation
@online{laderas2022,
author = {Laderas, Ted},
title = {RStudio {Conf} 2022 {Highlights:} {Days} 1 and 2},
date = {2022-07-05},
url = {https://laderast.github.io/posts/2022-08-04-RStudio-Conf/},
langid = {en}
}