“The goal of reproducible research is to tie specific instructions to data analysis and experimental data so that scholarship can be recreated, better understood and verified.” Max Kuhn, CRAN Task View: Reproducible Research

Our work exists on a spectrum of reproducibility

alt text

Peng 2011, Science 334(6060) pp. 1226-1227

In this module you will explore the use of software called Git to manage ‘versions’ of files. Similar to ‘track-changes’ in Microsoft Word, Git keeps track of any edits and makes it possible to track who made the change and when. Git (and other version control software) are most commonly used to manage collaboratively edited code, but it can keep track of any file.

Version Control

Tracking changes with version control


  • Eases collaboration
  • Can track changes in any file type (ideally plain text)
  • Can revert file to any point in its tracked history


  • Learning curve


alt text

  • Strong support for non-linear development: Rapid branching and merging, specific tools for visualizing and navigating a non-linear development history.
  • Distributed development: No central server needed, each user has a full copy
  • Efficient handling of large projects: Designed to manage the Linux OS
  • Cryptographic authentication of history: The ID of a particular version depends uponthe complete history. Once published, it is not possible to change the old versions without it being noticed.

Git Has Integrity

Everything checksummed before storage and then referred by checksum.

It’s impossible to change the contents of any file or directory without Git knowing. You can’t lose information in transit or get file corruption without Git being able to detect it.


A way of reducing digital information to a unique ID:

alt text

A 40-character hexadecimal SHA-1 hash: 24b9da6552252987aa493b52f8696cd6d3b00373

Git doesn’t care about filenames, extensions, etc. It’s the information that matters…

Git Tutorial: let’s get started

Your turn

Take 15 minutes or so at this site to walk through some basic git commands
alt text

The 3 states of files

staged, modified, committed

alt text

The important stuff is hidden in the .git folder.


GitHub is a web-based Git repository hosting service. It offers all of the distributed version control and source code management (SCM) functionality of Git as well as adding its own features. It provides access control and several collaboration features such as bug tracking, feature requests, task management, and wikis for every project. Wikipedia

There are other ways to use Git, you can host your own server or use another private company, such as BitBucket.

You can think of GitHub as part:

  • Server to back up your files
  • Website to share your files
  • Method to track changes to your files
  • Platform to collaboratively develop code (or other files)

Example: this course website is managed using Git & GitHub

alt text

Commit to GitHub from within RStudio


  1. Edit: make changes to a file in the repository you cloned above
  2. Stage: tell git which changes you want to commit
  3. Commit (with a message)
  4. Push: send the updated files to GitHub


alt text

Select which changed files (added, deleted, or edited) you want to commit.


alt text

Add a commit message and click commit.

Syncing (push)

alt text

Click the green arrow to sync with GitHub.

Git File Lifecycle

alt text

Git command line from RStudio

RStudio has limited functionality.

alt text

Git help

$ git help <verb>
$ git <verb> --help
$ man git-<verb>

For example, you can get the manpage help for the config command by running git help config

Git status

alt text

Similar to info in git tab in RStudio

Git config

git config shows you all the git configuration settings:

  • remote.origin.url (e.g. to connect to GitHub)


Branches used to develop features isolated from each other. alt text

Default: master branch. Use other branches for development/collaboration and merge them back upon completion.

Basic Branching

$ git checkout -b devel   # create new branch and switch to it

$ git checkout master  #switch back to master
$ git merge devel  #merge in changes from devel branch

But we won’t do much with branching in this course…

Git can do far more!

Check out the (free) book ProGIT

alt text

Or the cheatsheet.


Remember, the data and code are real, the products (tables, figures) are ephemeral…

Git and RMarkdown

Visualize .md on GitHub

Update the YAML header to keep the markdown file

From this:

title: "Untitled"
author: "Adam M. Wilson"
output: html_document

To this:

title: "Demo"
author: "Adam M. Wilson"
      keep_md: true

And click knit HTML to generate the output

Visualize example

alt text

Explore markdown<->Git

  1. Use File -> New File -> R Markdown to create a new markdown file.
  2. Use the Cheatsheet to add sections (# and ##) and some example narrative.
  3. Stage, Commit, Push!
  4. Make more changes then Stage, Commit, Push!
  5. Explore the markdown file on your GitHub website.

Motivations: Claerbout’s principle

“An article about computational result is advertising, not scholarship. The actual scholarship is the full software environment, code and data, that produced the result.” Claerbout and Karrenbach, Proceedings of the 62nd Annual International Meeting of the Society of Exploration Geophysics. 1992