Introduction to R and Biostatistics (2012 version): presentation

To follow my Introducing R and Biostatistics to first year LCG students (2012 version) post,  you can now find the presentation online from my site either in presentation format, in a single webpage format, or the raw Rmd file. To prove the point that publishing to RPubs is super easy, you can also find the single webpage format over there. I also like how you can comment and share in RPubs.

One of the challenges of giving a presentation to first year students is finding the balance between introducing them to cool things you are doing in your work and actually giving a talk that they can follow. I thought about this and ended dropping anything related to my work.

My presentation was split pretty much in two parts. First, I wanted to promote some philosophical discussion about what is statistics. Second, I gave a brief overview of what you can do with R. Or more exactly, what they should be able to learn to do even if they become wet biologists.

While planning this presentation, I knew that I wanted to give the new students a flavor of the three different currents in statistics. I aimed to improve my 2011 explanations now that I’m taking the Foundations of Statistical Inference course. I’m happy with the result and I think this is greatly due to Royall’s diagnostic test example.

Another key point that I wanted to emphasize was that RStudio is the way to go if you are new to R. It is very straightforward to use, plus it is nicely interegrated with knitr.

I decided to use R Markdown (Rmd) for the first time, after seeing how easy Markdown really is compared to using LaTeX and Beamer. However, when it got to doing the presentation I have to say that I was a bit dissapointed by how some things just break when using the R Markdown to Markdown to HTML presentation pipeline —using pandoc for the last step. For example, the math breaks when using mathml or mathjax at times (like after adding an iframe for a youtube video), so I had to use webtex which doesn’t look as nice.

If you are interested in the commands, I used the “Knit HTML” button in R Studio [equivalent to running from R: library(knitr); knit(“filename.Rmd”)] and then ran the following command:

pandoc -s -S —webtex -i -t dzslides intro_R_Biostat_LCG_2012.md -o intro_R_Biostat_LCG_2012_slides.html & open intro_R_Biostat_LCG_2012_slides.html

I was originally aiming to have a single Rmd file to produce an HTML presentation and a Beamer presentation. However, controlling the pictures in the Beamer output proved to be challenging. While I had a work around, the final problem was the math part. By the time I realized this it was too late —I just dropped the Beamer presentation. Googling, even the author of knitr acknowledges that the best input for PDF output is still LaTeX.

In the end, I’m happy that I got the HTML presentation done using R Markdown and briefly introduced it to the first year students. The basics of knitr are very easy to learn and I’m hoping that it got some of them curious enough to try it.

Next thing in line: prepare 5 questions for the students.