You should be using renv
summary
Dependence management is a reproducibility toolchain: initialize a project-specific library with renv::init, capture exact package versions and sources in renv.lock, and reconstitute environments with renv::restore. A real incident shows installing dtplyr silently upgrades dplyr and breaks a dashboard that processed ~80,000 census tracts. renv keeps a shared package cache so subsequent installs are fast, reports status about missing packages, and lets projects migrate incrementally rather than forcing a simultaneous rewrite.
AI-written summary · about
chapters (4)
related talks
20mPractical {renv}
59mEmbracing multi-lingual data science
36mPipenv: The Future of Python Dependency Management
24mPipenv: New Package Dependency Management System for Python
26mSecure Modern and Simple Dependency Management with PDM and Friends
34mAdventures in pip land
26mDependency Management in Python
26mRepeatable Data Setup for Repeatable Science using Julia