RStudio.Cloud: Organizing Research with RMarkdown (1 Hour)

  • Due No Due Date
  • Points 0

Write a reproducible article with rticles

RMarkdown is an especially powerful tool for integrating the data analysis that produces your results into your paper. Sharing your work with integrated statistical code makes it easier for others to understand and reproduce your work. Replicability is a hallmark of good science, but also highly useful to those of you who may be like me and juggling many projects at once. Well organized statistical programs enhance your own confidence in your work and lowers your switching costs when transitioning between projects.

For this exercise, head to our RStudio.Cloud workspace and launch the Rticles Workspace. Start with the _readme file.

First time using RStudio.Cloud? Start with this Intro and Tour of RMarkdown video Links to an external site. first and join our RStudio.Cloud workspace using this link Links to an external site. to work through the introductory exercises as you watch.

This Rticles exercise is adapted from a full online course (open-source) available in full here: https://annakrystalli.me/rrresearch/about.html Links to an external site.

In the rticles tutorial, you learn about citr, which provides a nice menu-based way to navigate a bibtex file and add your citations. However, isn't part of the reason we're learning to write papers in R to get away from drop down menus? (the answer is yes, yes it is!) And citr has some vulnerabilities and dependencies that may make it less stable than other strategies, though the user interface is nice. I'd like to make sure you have another strategy in your toolkit in case citr ends up being more trouble than it is worth for you.

This post Links to an external site. provides a nice introduction to using Zotero - which is a reference manager like EndNote. It will:

  • Allow you to painlessly clip in references to your library using browser plugs in, for example from a Google Scholar search
  • It will allow you to export a .bib file containing your library for reference in RMarkdown documents
  • It also has a style repository that you can go to for custom csl files to place in your YAML header. This means whatever crazy specific way your target journal wants your citations, RMarkdown will update them to fit that style when knit with the csl in the YAML (provided your bib file is in good shape)
  • It plays nicely with RMarkdown's Visual editor Links to an external site. allowing you to bypass citr and still get some handy menu help, plus it will work if you are on RStudio Server or Connect and don't have Zotero installed locally (it uses the Web API)

No solution is totally tinker free, but I hope adding this option to your toolkit will offer you more flexibility - or at least give you a new target for frustration :)

Zotero is free to use. Go here to set up an account Links to an external site., here to download Links to an external site. the desktop version and browser plugins, and here to access a quick start guide. Links to an external site.

Looking for some low effort version control? Below is a yaml I wrote to use for both analysis files and my full manuscripts. Every time it knits, it produces a new document so old versions are automatically preserved when I knit and can be recovered.

To use this, you need to edit as follows:

  • Fill in your title and name per usual.
  • Create the following file structure:
    • House all project files in one main folder.
    • Save Rmd files in a subfolder (I name mine Prog)
    • Create another subfolder called Notes (this is where the knit files will go). If you name yours something else, edit the yaml out_dir
    • I also create subfolders for original data (Orig) that I never write over and cleaned up data (Data) that gets written over when files from Prog are run.
  • Edit filename to replace '1_name_of_rmd' with whatever you are naming the rmd file itself. (For example '1_datacleaning'). This ensures the Notes folder names will match the name of the Rmd that creates them.
  • The yaml uses relative paths because it has to but while we're on the topic, the here package is great for the main body of your rmd and the rest of your workflow (learn how Links to an external site.)

---
title: "Clever Title"
author: "Your Name Here"
date: "`r format(Sys.time(), '%d %B, %Y, %H:%M')`"
knit: (function(inputFile, encoding) { 
      out_dir <- "../Notes";
      filename <- paste0('1_name_of_rmd',Sys.Date(),'.html');
      rmarkdown::render(inputFile,
                        encoding=encoding, 
                        output_file=file.path(dirname(inputFile), out_dir, filename)) })
output: html_document

---

 

 

 

 

0