Data without software are just numbers

A growing trend in producing academic research is abiding by the FAIR principles, which state that produced research data be findable, accessible, interoperable, and re-usable. These, in theory, lend to the important concept of reproducibility. However, what about the software used to generate the data? Often that software is a home-grown solution, and the software and its developers are rarely cited in academic research. This does not lend to reproducibility. As such, researchers such as Davenport et al. have written on the topic of improving research output reproducibility by addressing good software development and citation practices. In their 2020 paper published in Data Science Journal, they present a brief essay on the topic, offering background and suggestions to researchers on how to improve research software development, use, and citation. They conclude that “[e]ncouraging the use of modern methods and professional training will improve the quality of research software” and by extension the reproducibility of research results themselves.

Please to read the entire article.