R Project Setup

Initialize a new R data science project with best practices. This command uses the r-data-architect agent to create a well-structured, reproducible project foundation.

Project Structure Created

project_name/
├── R/                      # R functions
│   └── utils.R
├── data/
│   ├── raw/                # Original, immutable data
│   └── processed/          # Cleaned data
├── output/
│   ├── figures/            # Plots and visualizations
│   ├── tables/             # Summary tables
│   └── models/             # Saved model objects
├── reports/                # Quarto/RMarkdown reports
│   └── analysis.qmd
├── tests/                  # testthat tests
│   └── testthat/
├── _targets.R              # targets pipeline (optional)
├── .gitignore              # Git ignore patterns
├── .Rprofile               # Project-specific settings
├── renv.lock               # Dependency lock file
├── project_name.Rproj      # RStudio project
├── README.md               # Project documentation
└── DESCRIPTION             # Project metadata

Usage

/r-project-setup [project_name] [template] [options]

Parameters

• project_name: Name for the new project
• template: One of analysis, package, shiny, clinical (default: analysis)
• options: Comma-separated options (see below)

Templates

analysis: Standard data analysis project

• Focus on EDA and statistical analysis
• Includes Quarto report template
• Sets up targets pipeline

package: R package development

• Standard R package structure
• Includes testthat setup
• pkgdown configuration
• GitHub Actions workflows

shiny: Shiny application

• golem or rhino structure (configurable)
• Modular app organization
• Testing infrastructure

clinical: Clinical trial analysis

• Pharmaverse-ready structure
• Validation documentation templates
• TLF folder organization

Options

• renv: Initialize renv for dependency management (default: yes)
• targets: Set up targets pipeline (default: yes for analysis)
• git: Initialize git repository (default: yes)
• github: Create GitHub repo structure (default: yes)
• docker: Include Dockerfile (default: no)
• ci: Include CI/CD workflows (default: yes)

Example

/r-project-setup customer_churn analysis renv,targets,git

This creates an analysis project called "customer_churn" with renv, targets, and git.

Files Created

Core Files

README.md

# Project Name

## Overview
Brief description of the project.

## Setup
Instructions for setting up the environment.

## Usage
How to run the analysis.

## Structure
Description of project organization.

DESCRIPTION

Type: project
Package: project_name
Title: Project Title
Version: 0.1.0
Authors@R: person("Name", "email@example.com", role = c("aut", "cre"))
Description: Project description.
License: MIT

R Files

R/utils.R

• Common utility functions
• Data loading helpers
• Project-specific functions

Configuration

_targets.R (if targets enabled)

library(targets)

tar_option_set(
  packages = c("tidyverse", "tidymodels"),
  format = "qs"
)

list(
  tar_target(raw_data, read_data("data/raw/data.csv")),
  tar_target(clean_data, clean_data(raw_data)),
  tar_target(model, fit_model(clean_data)),
  tar_target(report, render_report(model))
)

.Rprofile

source("renv/activate.R")

options(
  tidyverse.quiet = TRUE,
  scipen = 999
)

GitHub Actions

.github/workflows/check.yml

• R CMD check
• Test coverage
• lint checking

.github/workflows/targets.yml

• Pipeline execution
• Caching

Dependencies Installed

Standard analysis project includes:

• tidyverse
• tidymodels
• targets
• quarto
• gt, gtsummary
• here
• janitor

Post-Setup Steps

After creation, the following manual steps are recommended:

1. Review and edit README.md with project-specific details
2. Add your data to data/raw/
3. Update _targets.R with your pipeline
4. Configure renv with additional packages: renv::install("package")
5. Push to GitHub if using version control

Notes

• Project is created in a new subfolder
• All paths use the here package for portability
• Template files include helpful comments
• Configuration follows tidyverse best practices