Best books and online classes for learning R

Happy New Year!

How about a resolution to learn something new in 2018?

Under the time-honored medical school principle of “See One, Do One, Teach One,” I am preparing to help medical students use the R programming language. This post is to share what I’ve learned about learning R.

Set up your PC: download R and then download RStudio.

Advice from a friend who teaches machine learning at Harvard… “For a traditional procedural programmer”: read R Cookbook (O’Reilly); “For someone with low-level or Lisp knowledge”: read Advanced R (CRC).

[Somewhat tangentially, he recommended An Introduction to Statistical Learning, with Applications in R (James, et al; Springer), or The Elements of Statistical Learning (Springer), which can be used to awe friends with all of the equations and graphs.]

If you’d rather watch lectures and take short quizzes, start with the free edX course Data Science: R Basics. The next step on edX is Statistics and R, part of a seven-course series.

Separately, I’m not sure that I love R so far. “Like APL without the special keyboard” seems like a fair description. All kinds of magic happen with just a character or two and I worry that code won’t be readable, maybe even by the original author!

7 thoughts on “Best books and online classes for learning R

  1. Anything by Hadley Wickham is great. There is also a book out there along the lines of “R for Biologists” and MOOCs like the one from Hopkins. Finally,R syntax is quirky and Python is very strong in data analysis now (Pandas, NumPy, etc). Seems less of a stretch for a non Programmer audience (but R does graph better)

  2. Agree with TimB. Why torture fresh minds with R?

    Virtually the entire ML community has switched to Python in the past 5 years. I still see some Matlab code in new papers on the arXiv, but >90% is Python.

    (Full disclosure: as an employer of ML researchers, I’ve long insisted that everyone work in Python. It is far easier to learn, read, test, and repeat results.)

  3. Tim, John: Thanks for the Python idea. As with everything else on Planet Earth these days, it is not my decision to make! (i.e., I’m helping with the course rather than leading it) Also, I think the students prepared for this class during the fall of 2017 by studying R.

  4. A 24 year old language is more interesting than the modern languages, because they weren’t created to drive the profits of 1 company as much as modern languages & they’re a bit of retro computing.

  5. Both are good but I still wonder about the target audience Harvard, Hopkins, Duke, Microsoft and likely more all have MOOCs based in R Best of luck!

  6. the way I teach R to students is this:

    if you cannot learn R by yourselves by the end of the week, that might just be enough info on whether you are cut for this or not.

    Really, teaching R? needing help to learn that? Just avoid RStudio, it has a terrible graphical engine that makes printing graphs with loads of point tediously slow.

  7. I see a lot of pre-med and med schoolers in the coffee shops I frequent. What strikes me is that their primary learning tools are game-ified and manga-fied, and are on their computers. Perhaps you can go old-school and make an R coloring book, like the anatomy coloring books used when I was still a young man.

Comments are closed.