In-depth topics

Chapter-length notes on individual topics that come up in the templates but don’t fit cleanly into a README. Each file is a standalone reference — read whichever one you need, in any order. Linked from the hub README.

Chapters

The list will grow as topics are written. Linked entries below are live; unlinked entries are placeholders for content I plan to add.

  • Project structure for research — the three storage locations (GitHub, local clone, cloud sync), why code and data must be separate, why the local clone shouldn’t be inside Dropbox, the recommended folder layout, file-naming conventions, and the LaTeX-vs-Word manuscript decision.
  • Git and GitHub for research projects — Git vs. GitHub, the everyday commit/push/pull workflow, template repositories, branching for R&R revisions, tagging paper versions, .gitignore essentials, and a glossary.
  • Setting up your IDE — what an IDE is and the popular ones, why an IDE beats Jupyter notebooks for reproducible research code, how IDEs treat a folder as a project, the source-control panels in RStudio and VS Code, per-editor setup notes, ligature fonts, and AI-assistant integration.
  • Python virtual environments with venv and uv — what a Python virtual environment is, why you need one per project, the messy Python tool landscape (venv / pip / pip-tools / pyenv / pipx / poetry / conda) and why the templates standardize on uv, the pyproject.toml + uv.lock model, and the five uv commands you’ll actually use day-to-day.
  • Environment variables and the .env file — what an environment variable is, why the templates use .env instead of hardcoded paths, how each of R / Python / Stata / SAS reads it, credentials in keyring vs. .env, common gotchas.
  • About AGENTS.md — what the AGENTS.md file in each repo is, the cross-tool standard behind it, how it interacts with CLAUDE.md, and why you might want one in your own projects.
  • Replication packages and what data editors expect — what the major data and code sharing policies actually require, the split between disclosure and verification (and why accounting sits on the wrong side of it), the Social Science Data Editors’ README template, how licensed WRDS data is handled, and which artifacts the templates generate for you.
  • SAS macros and batch_run_sas — MACROS.sas, a working SAS script (002-merge-fdp-data.sas from the consensus repo), and a walkthrough of using %load_env + batch_run_sas() for JAR-style logs.

Table of contents