R to Rmd

This document presents usual steps to convert a R file to a Rmd file.

1 - Open RStudio

2 - Open your R file (with .R extension)

It may contain :

• call to libraries
• comments
• code to load data
• code to modify data
• code to make figure

(run all the lines to see)

3 - Open a Rmd file

using File > New File > Rmarkdown

Two options :

• Create empty document to start from scratch
• Name and save the document to start from a pre-filled Rmarkdown (useful !)

A Rmarkdown file is structured with :

• header : many options are possible to customize style
• chunks to write with programming language
• empty space between chunks to write in natural language

Note : To insert chunk, use Code > Insert chunk, or the shortcut (CTRL + ALT + I)

There are many options to customize chunks. A guide can be found at https://bookdown.org/yihui/rmarkdown/.

4 - First rendering

4.1 - Copy paste all the R file content into a Rmd chunk

4.2 - Knit it !

to see a first rendering, by pressing the Knit button

Notes :

• It creates a new file with html extension, with the same file name as the Rmd
• Warning : If output file already exists, it will be overwritten

Observations :

• Good traceability
• All figures are there
• One single output file : easily shareable

5 - Second rendering

5.1 - Split chunks and comments

5.2 - Run each chunk

We can :

• appropriate the code by adding comments
• explain figures

5.3 - Knit it !

6 - Third rendering

6.1 - Add formatting

Comments are formatted in markdown (R-markdown !) :

• add titles using # (first level), ## (second level) … up to 6 !
• make italic text : *italic text* or (_italic text_)
• make bold text by wrapping it within two asterisks : **bold text**
• and more : https://www.markdownguide.org/cheat-sheet/

6.2 - Knit it !

7 - Fourth rendering

7.1 - Programming good pratice

In order of appearance :

• meaningful title
• current date : knitting date, to remember when the analysis was done
• explain goal of the document
• add .libPaths() to track the location of libraries
• load environment : libraries, global variables (eg. paths)… Group everything at the beginning of the script to see what will be needed !
• load data
• track libraries of origin for each function using package::function syntax instead of function, because some packages name their functions the same (eg. filter function)
• summarize the environment with sessionInfo() as the last line, in the last chunk

7.2 - Knit it !

Waouh !

8 - In practice

In practice, we start with a pre-filled Rmarkdown document, and add headers, codes, comments and formatting while progressing in the analysis. Programming best pratices should be applied while working, and not in the end…