Local Development Setup

R, RStudio, and Quarto on your machine

This guide covers setting up a local development environment. Skip this section if you’re using a Longleaf-only workflow.

NoteWhen to Use Local Development

Local machines are fine for:

  • Writing and editing code
  • Small-scale testing
  • Manuscript writing

For production analyses (simulations, calibrations, long-running jobs), use Longleaf.

Install R

Platform Method
macOS Download from CRAN or brew install r
Linux (Ubuntu/Debian) sudo apt install r-base r-base-dev
Linux (Fedora) sudo dnf install R
Windows Download from CRAN

Install RStudio (Optional)

Download from posit.co

RStudio provides an integrated development environment with:

  • Script editor with syntax highlighting
  • Console for interactive R
  • Environment browser
  • Git integration
  • Quarto support

Essential Packages

# Core packages
install.packages(c(
  "data.table",
  "targets",
  "tarchetypes",
  "crew",
  "here",
  "yaml"
))

# For clinical trial work
install.packages(c(
  "survival",
  "DiceKriging",
  "lhs"
))

# For cluster computing (if using crew.cluster)
install.packages("crew.cluster")

Local Quarto Setup

Download from quarto.org/docs/get-started

Platform Method
macOS Download installer or brew install --cask quarto
Linux Download .deb or .tar.gz from website
Windows Download installer from website 
# Verify installation
quarto --version
quarto check

LaTeX for PDF Output

Quarto can install TinyTeX automatically:

quarto install tinytex

This provides a minimal LaTeX distribution sufficient for most PDF rendering. If you need additional LaTeX packages:

# In R
tinytex::tlmgr_install("package-name")

Platform-Specific Notes

Windows

WarningWindows Limitations

Some packages and workflows have issues on Windows:

  • Path separators: Use here::here() or forward slashes
  • Line endings: Configure Git to handle (git config --global core.autocrlf true)
  • Some packages: May fail to compile (use pre-built binaries when available)
  • Parallel computing: mclapply() doesn’t work; use parLapply() or future
  • Shell scripts: Won’t run natively; use Git Bash or WSL

Recommendation: For serious work, use Longleaf or install WSL 2.

macOS

# Install Xcode command line tools (required for package compilation)
xcode-select --install

# If using Homebrew
brew install r
brew install --cask rstudio

The Xcode command line tools provide compilers needed to install R packages from source.

Linux

Most R packages compile smoothly on Linux. You may need development headers for some packages:

# Ubuntu/Debian
sudo apt install libcurl4-openssl-dev libssl-dev libxml2-dev

# Fedora
sudo dnf install libcurl-devel openssl-devel libxml2-devel

Local .Rprofile Template

Create or edit ~/.Rprofile on your local machine:

# ~/.Rprofile (local machine)

# Set CRAN mirror
options(repos = c(CRAN = "https://cloud.r-project.org"))

# Useful options
options(
  warnPartialMatchDollar = TRUE,
  warnPartialMatchArgs = TRUE
)

These options help catch common bugs:

  • warnPartialMatchDollar: Warns when using $ with partial column names
  • warnPartialMatchArgs: Warns when function arguments are partially matched

Verification

After setup, verify everything works:

R --version
Rscript -e "library(targets); print('OK')"
quarto --version
git --version
gh --version  # If you installed GitHub CLI

In R:

# Check essential packages load
library(data.table)
library(targets)
library(here)

# Check Quarto integration
quarto::quarto_version()

Syncing with Longleaf

If you develop locally but run on Longleaf:

  1. Use Git to sync code between machines
  2. Keep data on Longleaf only (use symlinks for local references)
  3. Test locally with small subsets
  4. Run full analyses on Longleaf

See Git & GitHub Setup for version control configuration.

Next: Git & GitHub Setup | Longleaf Setup