R Data Science Project Template

A minimal, opinionated project template for R-based data science analyses. Designed for individual or small-team projects where simplicity matters more than enterprise tooling.

Project Organisation

├── README.md                   <- The top-level README for this project
├── project-name.Rproj          <- RStudio project file; open this to launch the project
│
├── data
│   ├── raw                     <- Original, immutable data files
│   └── processed               <- Cleaned, transformed data ready for analysis
│
├── R                           <- Reusable R functions and utilities
│   └── utils.R                 <- Helper functions sourced by analysis scripts
│
├── analysis                    <- Numbered analysis scripts, run in sequence
│   ├── 01_data-cleaning.R      <- Data import, validation, and cleaning
│   ├── 02_eda.R                <- Exploratory data analysis
│   ├── 03_analysis.R           <- Primary analysis (modelling, clustering, etc.)
│   └── 04_supplementary.R      <- Additional analyses as needed
│
├── outputs                     <- Generated files from analysis scripts
│   ├── figures                 <- Plots and visualisations (.png, .pdf, etc.)
│   └── tables                  <- Summary tables and exported results (.csv, .xlsx)
│
└── report                      <- Final reporting documents
    └── final-report.Rmd        <- R Markdown report compiling results and narrative

Getting Started

Using the here Package

This template relies on the here package for relative path management. Combined with the .Rproj file, it eliminates the need for setwd() calls.

install.packages("here")

# Then in any script:
library(here)
data <- read.csv(here("data", "raw", "your_data.csv"))

Always open your project via the .Rproj file to ensure paths resolve correctly.

Adjusting .gitignore

Ensure you adjust the .gitignore file according to your project needs. By default, the /data/ folder is excluded from version control:

# exclude data from source control by default
/data/

You may want to track data if it's small and non-sensitive, or ensure it's excluded if it contains personally identifiable information or large files.

Environment Management (Optional)

For projects requiring strict reproducibility, initialise renv:

install.packages("renv")
renv::init()

# After installing packages:
renv::snapshot()

For coursework or exploratory analyses, this is typically unnecessary.

Workflow

1. Place raw data in data/raw/ — never modify these files directly
2. Run analysis scripts in order (01_, 02_, etc.)
3. Save processed datasets to data/processed/
4. Export figures and tables to outputs/
5. Compile final report from report/final-report.Rmd