By Cameron Cook
The Rebecca J. Holz series in Research Data Management is a monthly lecture series hosted during the spring and fall academic semesters. Research Data Services invites speakers from a variety of disciplines to talk about their research or involvement with data.
This November, Karl Broman, a professor in the Department of Biostatistics and Medical Informatics at the University of Wisconsin – Madison, gave a talk entitled “Reproducible Research”. You can find the slides and notes for his talk on his website.
Karl’s talk centered on steps that can be implemented over time to make one’s work fully reproducible. He began with an overview of the issues he faces in his work, highlighting common problems and misconceptions that many researchers have experienced. He also introduced the idea of what reproducibility means and how that idea differs from replicable research. The steps to reproducibility that are outlined in Karl’s talk are listed below, further explanation can be found in the notes attached to his archived slides.
1) Everything with a script – Everything you do, do in code.
2) Organize your data and code – Make your code or data meaningful to someone else.
3) Automate the process – Karl uses GNU Make for the process.
4) Turn your scripts into reproducible reports – Give a better picture of the data.
5) Turn repeated code into functions – Write better code!
6) Create a package/module – Don’t repeat yourself. Reuse and improve upon the code you have.
7) Use version control – Through tools like git/GitHub.
8) License your software – License your code so you can share it with others.