Intermediate Git for Biostatistics

A Three-Day Bridge from the boot camp to the Practicum

Author

The rgtlab Curriculum Project

Published

2026-07-28

Welcome

This is the online version of Intermediate Git for Biostatistics: A Three-Day Bridge by The rgtlab Curriculum Project. It is a short second course in version control, situated between the boot camp and the graduate sequence proper.

The companion volume Git and GitHub for Biostatistics: A One-Week Boot Camp takes a student from never having made a commit to collaborating through pull requests. This bridge picks up there. It teaches the intermediate skills a student reaches for within days of finishing the boot camp: committing a clean subset of a messy working directory, shelving work in progress, curating a branch’s history into something a reviewer can read, and recovering when an operation goes wrong. These are the skills that separate a tidy, reproducible project history from a serviceable but untrustworthy one.

The bridge is deliberately short, three days rather than five, at the same daily cadence: one hour of lecture, two hours of homework, no examinations. It is not a survey of everything Git can do. The infrastructure that surrounds version control in professional practice, namely continuous integration, containerisation, dependency pinning, and data-versioning tooling, is left to the Biostatistics Practicum, which this bridge feeds directly.

The organizing principle is unchanged from the boot camp: version control is a research-integrity tool. Where the boot camp taught a student to record history, this bridge teaches a student to curate and to recover it, so that the commit log remains a faithful and legible account of how an analysis was built.

The scope of each day:

  1. Selective staging and work in progress. Committing part of a file with patch-mode staging, discarding selective changes, and shelving unfinished work with the stash.
  2. Rewriting history. Amending, interactive rebase, cherry-picking, and the discipline that governs when rewriting is safe and when it is forbidden.
  3. Recovery and forensics. The reflog as a safety net, git bisect to locate the commit that changed a result, and git blame and the pickaxe to investigate history.

Position in the sequence

This bridge is a preparatory companion, alongside the two boot camps, to the graduate sequence:

License

This book is licensed under Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International.

Code samples are licensed under Creative Commons CC0 1.0 Universal, i.e. public domain.