Reproducible science - motivation and basic tools
Marie-Pierre Etienne
https://github.com/MarieEtienne/reproductibilite
2024-07-24
Motivations
Reproducible science
In their paper @alston2021beginner address the question of science reproducibility in ecology.
The key aspects
It’s a growing trend: more and more journals require some elements of reproducibility.
Motivations
Practically
Where is the correct version of the report?
Motivations
Practically
An explanation in images: vidéo by Ignasi Bartomeus.
Course objective
Introducing tools to facilitate step 2.
Overview
Code / paper writing
- git for
- keeping track of different versions of code
- collective work
- reviewing and improving code quality
- keeping track of different versions of code
Around the code
- Rmarkdown, Jupyter notebooks, Quarto to bring together code, results, and analysis.
- vignettes in R packages
- R help pages
- some scientific papers (Computo asks for Jupyter/Quarto submissions)
- dashboards …
- vignettes in R packages
Overview
Controlling software versions
Reproducibility also relies on stabilizing programming/library versions.
The container system Docker allows freezing a working software environment.
The winning trio of reproducible research with R
- git code versionning
- Rmarkdown markup language
- Docker Controlling software version (last 2 sessions with Theo Fabien)