R Markdown Tips

R Markdown repetition and a few tips

Let’s start with creating an r project again. To use this with an existing git repo, with can select “Existing dir”. You can also create an empty one and move your git repo in here later, as long as there is an .git folder RStudio will show you the context menu for git.

Next, let’s create an R Notebook. We could have a create an R Markdown document also, but the notebook offers a few conveniences. Mostly that it has a preview option html_notebook that renders the notebook to HTML in its curent state. In contrast, knitting the notebook to HTML via html_document will run all cells so this takes longer.

Note that it is important to knit to HTML before sharing so that you are sure everything works from scratch. This is the same reason we should do “Run all” in Jupyter Lab before sharing and why we don’t want to store our R workspace sessions. We need to make sure that someone new can run this from it’s current state. Another useful tool for this in R is to use devtools::session_info() at the top or bottom of your document (I put it at the end of the chunk where I load libraries) to ensure that you have included information about the versions of the packages you are running so that someone else can use the same version. There are more robust ways of version control that we will get into later in MDS, but this is a good minimum measure that is easy for you to get into the habit of doing.

A couple of features that are good to know in addition to those we learnt last time, are block commenting and automatic code reformatting. If I type a few lines where I for example forget to add whitespace around an operator or assignment, going to Code -> Reformat code ( Ctrl + Shift + A on Windows/Linux or ⌘ + Shift + A on a Mac ) will fix this automatically for all highlighted lines. If I want to toogle commenting for some lines, I can click Code -> Comment/Uncomment line ( Ctrl + Shift + C on Windows/Linux or ⌘ + Shift + C on a Mac ), instead of manually adding # in front of each line.

One final tip is the use of the here package for file paths. We have already solved the part of setting working directory by creating an R proj. If you only plant on using R Markdown file, you would be fine writing relative paths (e.g. ../data/cars.csv) the same way you would write them in Python because they look relative their own location. However, if you also need to run something from a script or the console, note that the working directory path will now be used as the current directory, rather than the directoy of the script so the same relative path will not work (you would need data/cars.csv instead). here solves this by allowing you to type here::here('data', 'cars.csv') from wherever you are which also makes sure that file paths work across operating systems (more info on here here.

R Markdown YAML header

The YAML header (also called the “front matter”) is where we can specify metadata about our project. It is delimited by two --- (three hyphens) and we create a new notebook, it looks like this:

---
title: "R Notebook"
output: html_notebook
---

In YAML, data is stored as a key: value pair, just like a Python dictionary. We can add new values, for example the author name and the date.

---
title: "R Notebook"
output: html_notebook
author: Joel Ostblom
date: 2020-09-23
---

R code can be evaluated inside the YAML header, so if we wanted the date to be updated every time we stitch the document, we could instead write date: `r Sys.Date()`. Other useful options include the ones for numbering headings, adding a table of contents, and placing the table of contents on the side of the document. Since these are options to the output document, they are indented under that section with two or four spaces:

---
output:
  html_notebook:
    toc: yes
    toc_float: yes
    number_sections: yes
---

Another useful option is to fold away your code, but still having it available for view if someone desires to see it.

---
output:
  html_document:
    code_folding: hide
---