Become an R Package Developer!
👋 Hei!
I’ve never been in Bergen
😭
But… I’ve been to Norway two times when I was an intern in Gothenburg, Sweden!
Me in front of royal Palace in Oslo, 2010
Me in (near?) Evenstad, 2010. Picture by Anne-Sophie Bonnet-Lebrun
My R package development creds
I really ❤️ R package development
-
Volunteer editor for rOpenSci Software Peer Review
-
At work, maintenance of rOpenSci dev guide
-
Created the R-hub blog (👋 guest author Julia!)
-
Working on the HTTP testing in R book
Why develop an R package?
Easiest way to share code/data/R Markdown templates… with
-
(future) you,
-
the collaborators you know,
-
and the collaborators you don’t.
Why learning about package development?
Jon Calder’s very good wording
-
To share code (and data)
-
To leverage existing tooling
-
To contribute to other packages
Who can write a package? YOU!
Susan Johnston’s wise words.
-
Can you open R or RStudio?
-
Can you install a package?
-
Have you ever written a function in R?
-
Could you learn how to write a function in R?
-
You can write a package in R
What is a package?
Pour réduire ses craintes, il faut se dire que ce n’est ni plus ni moins qu’un dossier organisé d’une manière contrainte.
To be less afraid you have to tell yourself that it’s simply a folder organized in a constrained way.
Automation!
Meme image by Allie Brosh
Automating… How?
Remember Clippy?
Automating… How?
Get to know an actually useful Clippy, {usethis}!
usethis and co
The usethis package implements this important principle: Automate that which can be automated. Your computer was literally born to implement rote-but-fussy stuff for you. https://t.co/OfIUzSBlc7
— Jenny Bryan (@JennyBryan) November 28, 2017
Website tour
Slides, notes from the demo, further resources, comments!
Time for a demo 🚀
Notes on the course website
Also refer to the Whole game chapter of the R packages book by Hadley Wickham and Jenny Bryan
Back from the demo
-
devtools::load_all()(and then install) -
{usethis}for all the things.
R CMD check (devtools::check())
Picture by Anne-Sophie Bonnet-Lebrun
So what’s really hard?
-
Writing good code
-
Writing a good interface
-
Writing good docs
More with packages
-
distributing data
usethis::use_data() -
packaging a Shiny app (look for examples, and golemverse)
-
Reproducible analyses, research compendium
Less with packages
-
If developing a package for wider distribution, check it does not exist yet.
-
Miles McBain’s post “Project as an R package: An okay idea”
Questions?
Write them in the Zoom chat, or as a comment in the course website, or as an issue in the course repository.
Time for a break! ☕
05:00