Reproducible research, Rnw, Rmd, makefiles and more


Global Moderator
Hi Everyone,

I'm starting the thread so we can exchange experience on reproducible research methods in the 21st century. Using tools such as R, knitr, latex, pandoc and others, we can easily create a methodology where the entire writing and analysis of a paper is entirely coded and reproducible.

I've created a template manuscript here on which I will build with knowledge gained, until it's the perfect manuscript skeleton: here

Please share your opinions on what makes the perfect academic manuscript skeleton/template. What should we use, which features to add. Which tool-boxes should we absolutely we using and why?

If many contributors have ideas or add to the skeleton, then I will make this a TS-github project. I'm sure we can make something very useful together. I know trinker has some ideas on what needs to be included in something like this.

Personally I'm writing my papers in .Rnw format, but as many of my co-authors are still using word, I was thinking about writing them in .Rmd. Which is a good compromise as pandoc can convert it quickly to all formats (see my skeleton on github).

I've encountered a few problems though:

1. bibliography citations and references are tricky
2. captions for figures and tables are doable but it's nothing like in latex
3. references to a figure or table are a pain
4. How do you handle double space, a requirement for draft manuscripts

Maybe you guys have already solved these things, maybe you have your own problems. No matter, lets solve them here.

Last edited: