Create R Research Project

A specialized sub-skill for scaffolding new R research projects with reproducible structure and best practices.

Overview

This sub-skill creates a complete, production-ready R project structure suitable for scientific research, statistical analysis, bioinformatics, and data visualization.

Use this sub-skill when the user wants to:

• Create a new R project from scratch
• Set up a reproducible research workflow
• Initialize Git for version control
• Generate analysis template scripts

---

What This Sub-Skill Does

When invoked, this sub-skill will:

1. Create directory structure

   • data/raw/ - Original, immutable data files
   • data/processed/ - Cleaned, transformed data
   • scripts/ - Analysis and processing scripts
   • results/figures/ - Plots and visualizations
   • results/tables/ - Summary tables
   • results/models/ - Saved model objects (.rds files)
   • reports/ - R Markdown/Quarto documents

2. Initialize version control

   • Create .gitignore with R-specific patterns
   • Initialize Git repository (if requested)
   • Create initial commit with project structure

3. Set up package management

   • Initialize renv for reproducible R environments
   • Create renv.lock file
   • Generate .Rprofile for automatic renv activation

4. Create RStudio project

   • Generate .Rproj file for RStudio integration
   • Configure project options

5. Generate template files

   • Analysis script template (scripts/01_analysis.R)
   • R Markdown report template (reports/report.Rmd)
   • README.md with project documentation
   • .gitignore for R projects

---

Example User Requests

• "Create a new R project for genomics analysis"
• "Set up an R research project with Git"
• "Initialize an R project for differential expression analysis"
• "Create an R workflow for statistical modeling"

---

Generated Templates

Analysis Script (scripts/01_analysis.R)

# Load libraries
library(tidyverse)
# Add required packages

# Load data
raw_data <- read_csv("data/raw/input.csv")

# Process data
processed_data <- raw_data %>%
  # Add processing steps
  mutate()

# Save processed data
write_csv(processed_data, "data/processed/cleaned.csv")

# Analysis
results <- processed_data %>%
  # Add analysis code
  summarize()

# Save results
write_csv(results, "results/tables/results.csv")

# Generate plots
g <- ggplot(processed_data, aes()) +
  geom_point() +
  theme_minimal()

ggsave("results/figures/plot1.pdf", g, width = 6, height = 4)

R Markdown Report (reports/report.Rmd)

---
title: "Analysis Report"
output: html_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(tidyverse)

Methods

Describe your analysis methods here.

Results

# Load and display results
results <- read_csv("results/tables/results.csv")
results

Figures

# Include generated figures
knitr::include_graphics("results/figures/plot1.pdf")

---

## Parameters

| Parameter | Description | Default |
|-----------|-------------|---------|
| `project_name` | Name of the project directory | Required |
| `init_git` | Initialize Git repository | `false` |
| `create_rproj` | Create RStudio project file | `true` |
| `init_renv` | Initialize renv environment | `true` |
| `project_type` | Type of project (general, bioinformatics, statistics) | `general` |

---

## Notes

- Project follows standard R research project conventions
- `renv.lock` ensures reproducible package versions
- `.gitignore` excludes sensitive data and compiled files
- All scripts use relative paths for portability

---

## Related Sub-Skills

- **run-analysis**: Execute analyses in created projects
- **debug-env**: Set up and troubleshoot R environments