R Markdown

Abstract

Authoring executable documents with R Markdown

R Markdown is a popular format for reproducible research. This article provides a guide to how to write, and collaborate on, a reproducible research article using R Markdown and Stencila. As well as describing the basics of embedding R code in your article, it covers how to include meta-data (such as author names and affiliations), citations and references. Finally, we discuss how you can preview your article and convert it to other formats using Stencila.

Introduction

R Markdown is a popular format for writing reproducible documents. Markdown is a simple text format, which, in the words of it's creator, is "intended to be as easy-to-read and easy-to-write as is feasible". R Markdown extends Markdown with some custom syntax for embedding executable R code.

There are many resources available for how to use R Markdown for reproducible research, including:

Code chunks

The basics

At the heart of R Markdown are "code chunks". Code chunks can produce outputs, or they can be used to execute arbitrary R code. For example, here we assign a variable called randoms which will be an array of 1000 random numbers sampled from a normal distribution:

randoms <- rnorm(1000)

When you assign a variable in a code chunk you can reuse it in another chunk, later in the document. For example, we can plot the frequency distribution of the random numbers we assigned to the randoms variable earlier:

hist(randoms, breaks=30, col="grey", main="")

Adding figure and table captions

R Markdown allows you to specify "options" for code chunks. One such option is fig.cap which allows you to specify a figure caption:

plot(randoms)

My figure

That is OK for short captions, but when you have longer captions you will probably want to use Bookdown style text references e.g.

plot(randoms)

This is a slightly longer caption for my figure including some strong emphasis.

Bookdown style text references are good for longer, single paragraph captions. However, for more structured captions having a title and paragraph as is often found in academic journals we suggest that you use a code chunk block extension. e.g.

plot(randoms)

The title of my plot

A paragraph for my figure including some strong emphasis.

Inline code chunks

In R Markdown you can also inline "inline" code chunks within paragraphs and other text content. For example, here is the mean of the randoms variable that we assigned earlier: mean(randoms) . In Stencila, we call these CodeExpressions, because they are intended to display a calculated value, and shouldn't be used for statements, such as assigning a variable.