Aiming for reproducibility of research results helps the community digest and check previous research. It also helps individual researchers achieve a cleaner workflow. Tools like Rmarkdown, knitr and Sweave are excellent, but partly obscure the full power of LaTeX and a clean separation between analysis/coding and paper/presentation writing.
The working example presented here provides an alternative. The main idea is that computational results are stored (using R or any other programming language) in CSV files and read into LaTeX. We provide an example of storing data in CSV files and convenience functions to read data from these CSV files into LaTeX in order to produce single-valued results as well as large tables of results.
More information is provided in the PDF file paper.pdf and its LaTeX-source paper.tex. The
R code that does the (dummy) data analysis is in R_code.r. If you change anything relevant
for the final paper.pdf, e.g., in paper.tex or R_code.r, you can use make followed by
make clean to rebuilt the whole paper.