Standards for reproducible R analysis pipelines. Use when setting up renv, targets, R Markdown, Quarto, or managing computational reproducibility.
renv)renv::init() at project start. Commit renv.lock to git.renv::snapshot() after installing/updating packages.renv::restore() to recreate exact environment.renv/library/ — only renv.lock and renv/activate.R.targets)_targets.R with list(tar_target(...)) for each step.targets::tar_make() runs only outdated targets.targets::tar_visnetwork()# _targets.R
list(
tar_target(raw_data, load_data("data/mortality.csv")),
tar_target(clean_data, clean_and_prepare(raw_data)),
tar_target(cb, build_crossbasis(clean_data)),
tar_target(model, fit_dlnm(clean_data, cb)),
tar_target(pred, predict_dlnm(model, cb)),
tar_target(fig_surface, plot_surface(pred))
)
set.seed(12345) at script start (or per stochastic block).sessionInfo() or sessioninfo::session_info().DESCRIPTION.output: html_document or pdf_document.#| label:, #| echo:, #| cache: in Quarto format.