R Project Template
The one-stop shop for reproducible R

Spin up a clean R project
in one click.

Name it, configure it, download it. You get a best-practice project skeleton — numbered pipeline or a targets DAG, config.R, renv, Quarto report, Makefile, Docker, Dev Containers, CI, and tests — zipped and ready to extract. No setup script to run.

Build my project → See every option
100% in your browser Nothing uploaded Save, export & share your setup 25+ toggles, zero lock-in

1 · Configure

Start from a preset
Save & share this setup (auto-saved in your browser)
Identity
Type & packages
Core packages tidyverse-first, with vroom for fast file I/O
Coordinate systems
Reproducibility & tooling
Extras
Additional packages

Any other CRAN/GitHub packages to library() in config.R (and install via renv/Docker).

Custom config variables

Extra constants appended to config.R under # Custom ----.

Custom folders

Additional directories created in the project root (e.g. 04_docs/notes).

2 · Preview


        

        

          

        

        

          

        

        

          

        

        

          

        

        

          

        

      

    

  


  

    
A pipeline you can actually reproduce

    
Every project follows the same numbered flow — raw → clean → model → visualize → report.

    

      
01
Download
Pull raw data into 01_data/raw_data/ — never edited by hand.

      
→

      
02
Clean
Transform & tidy into 01_data/clean_data/.

      
→

      
03
Model
Analyse; save results to 03_outputs/tables/.

      
→

      
04
Visualize
Maps & figures to 03_outputs/figures/.

      
→

      
05
Report
One Quarto doc ties it all together.

    


    

      
here::here() paths
Robust relative paths that work no matter where a script is run from.

      
renv lockfile
Pin every package version so collaborators get the exact same environment.

      
Docker option
Pin the whole OS + R version with a rocker-based image for total reproducibility.

      
Seeded
set.seed() wired into config so stochastic steps are deterministic.

      
Fast, tidy I/O
Tidyverse-first, with vroom for fast reads/writes of delimited data.

      
testthat + lint
Optional test scaffold and an air/lintr config to keep code clean.

      
targets DAG
Optional _targets.R pipeline that skips steps whose inputs haven't changed.

      
Dev Containers
One-click, fully pinned VS Code / GitHub Codespaces environment.

      
Save & share setups
Your builder configuration autosaves, and can be exported or shared as a link.

      
Fully customizable
Custom packages, config variables, and folders — add anything the catalogue doesn't cover.

    


    

      
After you download

      
    
        
  1. Unzip into your projects folder.
    2. 
        
  2. Open project.Rproj in RStudio.
    3. 
        
  3. Run renv::restore() to install pinned packages.
    4. 
        
  4. Start with 02_scripts/01_download_data.R and work through the files in order.
    5. 
      

    

  


  

    
Every option, at a glance

    
A true one-stop shop — toggle any combination on or off, the preview and download update live.


    

      

        
Identity & type

        

          
Project name / author / email / ORCID / descriptionFill config.R, README, CITATION.cff, LICENSE, Zenodo metadata

          
Spatial projectAdds sf/terra/leaflet + CRS config

          
Core packages18+ catalogue packages incl. tidyverse, vroom, arrow, DBI, googledrive, gt, plotly, patchwork, cowplot

          
Additional packagesAny CRAN/GitHub package not in the catalogue

        

      

      

        
Reproducibility & tooling

        

          
Random seed / min R versionset.seed() + version badge & requirements

          
Numbering stylePadded / plain / alpha / upper / none prefixes on folders & scripts

          
LicenseMIT, Apache-2.0, GPL-3.0, CC-BY-4.0, or none

          
Pipeline runnerMakefile, run.R, or both

          
targets pipeline_targets.R DAG with caching — skips unchanged steps

          
renvPinned, restorable package versions

          
uvruvr.toml + uvr.lock via uvr sync; replaces renv

          
Quarto reportNumbered .qmd report in HTML/PDF/Word

          
Git + CI.gitignore + GitHub Actions pipeline run on every push

          
.Rprofile + .RenvironSane session defaults + environment variable template

        

      

      

        
Extras

        

          
DockerfilePinned rocker image, spatial system libs when needed

          
Dev ContainerOne-click VS Code / Codespaces environment

          
testthattests/ skeleton wired into Makefile / CI

          
lintr.lintr + .editorconfig

          
airair.toml, make style, CI format check

          
pre-commitGit hooks that lint & format before every commit

          
GitHub templatesIssue forms, PR template, Dependabot

          
CITATION.cffMachine-readable citation metadata

          
Zenodo archive.zenodo.json for an archival DOI

          
Data dictionaryTemplate CSV + a README for 01_data/

          
Community docsCONTRIBUTING.md + Code of Conduct

          
Session logCapture sessionInfo() on every run

          
Custom config variables / foldersAnything the template doesn't already cover

        

      

      

        
Save, export & share

        

          
AutosaveYour setup persists in this browser between visits

          
Copy linkShare a URL that reopens the builder with your exact setup

          
Export / Import JSONSave a setup to reuse across future projects

        

      

    

  


  

    
FAQ

    

      

        
Which packages handle reading and writing data?

        
The core set is tidyverse-first. vroom is included for fast, tidy reading and writing of delimited files (it returns a tibble and plays nicely with dplyr), and conflicted can be toggled on to make any ambiguous function calls explicit.

      

      

        
Makefile, run.R, or targets — which do I need?

        
make all and Rscript run.R always re-run every step. targets::tar_make() builds a dependency graph and only reruns steps whose inputs changed — pick it for anything beyond a quick script.

      

      

        
renv, Docker, or Dev Containers?

        
They stack, not compete: renv pins R packages, Docker pins the OS + R version too, and a Dev Container wraps Docker with a ready-to-code VS Code/Codespaces setup. Use whichever matches how much environment drift you're worried about.

      

      

        
Is anything ever uploaded?

        
No. The whole builder — form, preview, zip generation — runs client-side in your browser with JSZip. Export/share features write to your clipboard, disk, or the URL bar only.