R Markdown is a miracle for scientific communication and repeatable research

At CRCTU, we recently ran our bootcamp course for the first time, a programme of core technical training for newly recruited biostatisticians.

Clinical trials units are busy centres of research where collaboration is the default. Being able to visualise data and communicate technical details to others quickly and effectively is important. Being able to reproduce something you calculated last month is necessary. R Markdown addresses both of these problems by interweaving prose and code to produce rich, interactive documents in … whatever format you need.

Here is RStudio explaining better than I could:

R Markdown is super flexible. Some things I have created in it are:

• A complete journal article, including abstract, authors and affiliations, keywords, figures and tables with captions, cross-referencing, bibliography (link to source)
• Literally this exact webpage you are reading right now (link to source)
• Interactive tutorials at https://rstudio.cloud/project/454702. Look for the Run.R file to launch a tutorial. (link to source)

In each example, the source code is written 100% in R Markdown and then compiled to create the article PDF or the web page or the Word doc or whatever. To emphasise that point, I have provided a link to the source Rmd file in each instance.

The R users at CRCTU use R Markdown files to conduct trial analyses and produce DMC reports. We trained new statisticians in this area because we knew that it is not the kind of thing universities would teach in their MSc programmes. However, reproducible research is one of the core pillars of our bootcamp programme and we see R Markdown as the key technology in this area.

Do you train statisticians in this area at your institution? Do you want to? Are you a statistician that wishes they had access to this type of training? If so, feel free to get in contact and we could bring it to your CTU.

Also, please fill in our brief questionnaire on how CTUs train statisticians!

Thanks!