Reproducibility in Research

Julia Romanowska

22.11.2022

WHY BOTHER?

It’s all about us the data

if we don’t document, organize, keep track and inform about the changes:
- nobody else can help us
- nobody else can continue our work (bus factor)
- nobody will trust and use our method or cite our paper

What should we do?

1. organize

good directory structure
https://slides.djnavarro.net/project-structure
consistent file naming
machine readable and human friendly
Check this presentation by Jenny Bryan
structured datasets
- How to create good datasets, even in Excel
- How to share data with a statistician

2. document

README files are essential
maintain a change log
comment the code/scripts
- if using R, Rmarkdown/Quarto is highly recommended!
keep digital notebook/blog

3. track changes

with 1. and 2. tracking changes is easy
and/or use git