Preface

This volume is a one-week boot camp for incoming graduate students in biostatistics who arrive with no or minimal prior exposure to version control. It is the second half of the pre-program preparation; the companion volume R for Biostatistics covers the R language over an equivalent week. Where that book makes a student functional in R, this one makes a student’s work reproducible and shareable. Both are prerequisites that every subsequent volume in the sequence assumes.

The design constraint is severe: five lectures, ten hours of homework, and the student arrives at the academic program able to place any analysis under version control and collaborate on GitHub. The constraint is met by ruthless prioritization. Git is a large and idiosyncratic tool; the great majority of its surface area is omitted deliberately. What remains is the irreducible core that a practicing biostatistician uses every day.

Why version control belongs in a biostatistics curriculum

The case for teaching version control to statisticians is not a case about software engineering. It is a case about research integrity. Consider three failures that every data analyst has seen or committed:

  • A results table in a manuscript cannot be reproduced because the script was edited after the table was made, and no record of the earlier version survives.
  • A folder contains analysis.R, analysis_final.R, analysis_final_v2.R, and analysis_FINAL_use_this.R, and nobody can say which produced the submitted figures.
  • Two collaborators email a script back and forth, each edits it, and their changes silently overwrite one another.

Each of these is a version-control failure, and each threatens the reproducibility that is the ethical foundation of quantitative science. Git solves all three directly: it records every version of every file with an author, a timestamp, and a message; it makes the history navigable; and it merges concurrent edits rather than losing them. A biostatistician who internalises this treats the commit history as a lab notebook, which is precisely what it is.

What this book covers

Five days, one chapter each:

  1. Why version control, setup, first repository. What version control is and why it is a reproducibility tool; installing and configuring Git; the working directory, staging area, and repository; git init, git status, git add, git commit, git log; the .gitignore file.
  2. Recording history: commits, diffs, undoing. The commit as a snapshot; git diff and git show; reading and navigating history; writing commit messages that earn their place; undoing safely with git restore, git revert, and git reset; amending.
  3. Branching and merging. Branches as movable pointers; git branch, git switch, git merge; the feature-branch workflow; merge conflicts and how to resolve them without fear.
  4. GitHub: remotes and publishing. What GitHub is; authenticating with the gh command-line tool; git remote, git push, git pull, git fetch, git clone; creating repositories with gh repo create; READMEs and issues.
  5. Collaboration and reproducible research. The fork-and-pull-request workflow; opening and reviewing pull requests with gh pr; resolving remote conflicts; using Git through the RStudio IDE; .gitignore for R projects, renv, and data; the handoff to the Biostatistics Practicum volume.

Each chapter is approximately 1 hour of reading and worked examples plus 2 hours of homework problems with worked solutions provided.

What this book does not cover

The book deliberately omits, and points elsewhere for, nearly everything beyond the daily core:

  • Git internals (the object database, the reflog, plumbing commands), interactive rebasing, cherry-picking, submodules, and hooks. See the Biostatistics Practicum.
  • Continuous integration, automated testing, and GitHub Actions. See the Practicum and Advanced Statistical Computing in the Age of AI.
  • Containerised and dependency-pinned reproducibility (Docker, renv in depth). See the Practicum.
  • Regulatory version-control practice for clinical trials (CDISC deliverables, validated pipelines). See the Practicum and Applied Statistical Methods for Health Sciences Research.

The boot camp is intentionally narrow. The student finishes Day 5 able to work fluently in the everyday Git workflow, leaving the Practicum free to teach the infrastructure that surrounds it.

How this book is meant to be used

The five-day cadence is the spine. A student starting the week having never made a commit can finish the week collaborating on GitHub. The cadence assumes roughly 3 hours of focused work per day:

  • Hour 1: Read the chapter, running every command in your own terminal as you reach it. Git is learned by typing, not by reading.
  • Hours 2-3: Complete the homework problems. Solutions are at the end of each chapter; check yourself only after attempting each problem.

The commands in this book are shown for a Unix-style shell (macOS Terminal, Linux, or Git Bash on Windows). Where a step differs on Windows, the difference is noted.

Prerequisites

The book assumes:

  • A working laptop running macOS, Windows, or Linux.
  • Basic familiarity with the command line: opening a terminal, changing directories with cd, listing files with ls. A short refresher is provided on Day 1.
  • No prior exposure to Git, GitHub, or any version-control system.

Prior completion of R for Biostatistics is helpful, since the worked examples version R analysis scripts, but it is not strictly required; the Git mechanics are independent of the language of the files under version control.

Acknowledgments

This boot camp consolidates patterns developed over two decades of teaching reproducible workflows to incoming biostatistics students. Jenny Bryan’s Happy Git and GitHub for the useR and the Software Carpentry version-control lessons provide the technical scaffolding; the boot-camp pedagogy reflects experience with what incoming students actually need in their first week, as distinct from the full capabilities of the tool.