[BlindMath] Learning R

[Godfrey, Jonathan]

Hello all,

I get asked about learning R all the time; many of these questions come from blind users, but mostly they come from sighted people.

Most documentation produced in the last few years is created using R markdown so ends up being extremely readable HTML as against the loathsome pdf documents of old.

I did spend a lot of time making more accessible documents for my own students and those blind people who needed accessible resources. I haven't really found this as necessary as was the case up to about five years ago.

I do still have resources at
https://R-Resources.massey.ac.nz/lurnblind
but even those pages haven't seen much attention for close on two years now.

Perhaps the greatest issue facing blind users of R  today is the inaccessibility of  RStudio. This integrated development environment (IDE) really does make life so much easier for the (sighted) novice R user. Rather unfortunately, many of the tools promoted by the corporate entity that is RStudio are confused with the RStudio IDE. The tools that do matter are add-on packages that can be used without the RStudio IDE, and are therefore still useful to blind users.

I think that anyone wanting to learn R should do so now in the context that is closest to their home university and the chosen course of study being undertaken. I'd add that the responsibility for working out which resources are the best fit should fall on the home university, although there is plenty of good advice already floating around.

A particular concern is that the R eco-system is now so vast that a person could be an expert at doing some things without having the skills to undertake what others consider to be basic.

I'm helping construct a new introductory course at the moment. In fact there are two sets of staff in my department constructing two different courses. Our learning objectives are very similar, but the means by which we'll get students to those objectives are starting to look very different. The course I am helping with will use R to create graphs and summary tables  at the outset and so our students will get exposure to the R syntax and way of working from day 1; the other course will use R for the things Excel cannot do which probably means what some people would think of as the hardest stuff in an introductory course, being confidence intervals and hypothesis tests. I'd also argue that this means R will be used for the most boring aspects of the course.

Why does this matter to anyone looking for resources to help learn R?
Well, if the "welcome to R" type tutorial comes before understanding of the topics for which R is being used, then the learner must learn R and the statistical stuff at the same time. Failure to understand hypothesis testing isn't an R problem.

So, I must express some caution about the choice of tutorial. There are literally thousands of them to choose from.  I suggest looking for tutorials that are small in scope, or easily broken into modules.

Common useful tutorials would be:


  *   How to import data
  *   How to graph data
  *   How to summarise data using numeric summary measures
  *   How to tabulate results
  *   How to fit a simple regression line

Aside from the last of these, I've seen ways to present the other topics in many different orders. Lots of teachers go past the data import challenges by making sure all data used in a course is totally sanitised and ready for use. That's fine for some classroom scenarios, but it doesn't help the budding scientist who has just spent six months in the field collecting data.

There really are plenty of wonderful reference works now in accessible html. Take a look at R for Data Science for an example of what I mean.
https://r4ds.had.co.nz/

Jonathan