Reference links for learning track

[colinleach]

Base R has been around for a long time, so there are lots of learning materials from various sources (of variable quality).
It's not entirely clear what to link to in our syllabus docs: a contrast to Julia, where docs.julialang.org is definitive.

Things are easier for anyone using RStudio, which has a help tab that can search all installed packages for reference documentation.

The key source is CRAN, which contains a huge amount of stuff - both software and documentation. It's also quite confusing, so I need to get better at navigating around the site.

I've used a lot of links to the manual hosted by ETHZ (a Swiss Federal University in Zurich). It's complete, quite ugly (a nostalgia trip for those of us who used VT131 terminals linked to a VAX in the mid-80's), and has no search facility as far as I can see. A few other universities have mirrors, but ETHZ seems to be the big one.

RDocumentation seems to use the same source, but with more visual styling (good) and adverts (bad).

All the Tidyverse packages come with excellent documentation, created in the last few years. That's the easy case.

There are a few online books. Introduction to R can be useful, alongside the CRAN intro of the same name.

Hadley Wickham is the guiding force for modern R, authoring several of the most popular packages as well as several of the most popular textbooks. Rather wonderfully, some of his books have dual-track publishing, with O'Reilly selling the print version and a free online copy on his website. These two are useful to us:

• R for Date Science is fairly introductory, but a large tome with a lot of content.
• Advanced R contains stuff that is hard to find elsewhere, useful if we need to get into details.

I have print copies of both of those, and highly recommend them.

I'm sure there are other sources that I'm missing, so I expect this will be an evolving thread.

The immediate question (in PR #413) is whether RDocumentation is a preferred source. Exercism guidelines just say we should avoid any site requiring registration (there are thousands of those for R). Personally, I'm allergic to adverts, but others may be allergic to web pages that would display virtually unchanged in a bash terminal.

[depial]

RDocumentation seems to use the same source, but with more visual styling (good) and adverts (bad).

I went with this instead of ETH because of the visual styling. I didn't know about the adverts, possibly because I have an ad blocker? I do see a stuff related to DataCamp which appears to run (or sponsor) the site, but I felt those were unintrusive enough to be tolerable. If there's more, I rescind any support for the site.

[BethanyG]

Quickly chiming in with this resource from RStudio education. Free to read, but not sure how up-to-date it is: https://rstudio-education.github.io/hopr/

There is also this resource from rstudio. Again, it might very well double what you have already listed -- but since posit/rstudio seem to be very active in the space, it might be a good place to skim.

[colinleach]

not sure how up-to-date it is

2014, but as R tends to add new stuff in packages while keeping Base stable, that's probably OK.

posit/rstudio seem to be very active in the space

Extremely active! Both in open-source and paid stuff. Hadley Wickham is now Chief Scientist there, so driving R modernization is his job as well as his hobby.

Like so many of the R creators, Hadley is a Kiwi from the University of Auckland. Good things can come from small, remote countries.

[colinleach]

Something I only just learned about (from Wikipedia):

In November 2023, Wes McKinney, creator of the Python package pandas, joined Posit as a principal architect. He was hired to advocate for the needs of the Python data ecosystem at Posit.

The name change from RStudio to Posit was also part of the language-neutral policy. I guess creating Shiny Python is an early example of this.