I scraped papers from PubMed published in JPR that mention a GitHub URL in the abstract (unfortunately it seems that even for OA papers the full text is not easily accessible) and calculated a few statistics. Sample size is 89 papers.
Many don't have a license. Otherwise various open-source licenses are used.
By definition unfortunately the majority of closed source software won't be included in this analysis because they won't list a GitHub URL in their papers.
GitHub stars don't correlate with paper citations. Nor do the number of contributors.
Let me know if this is interesting and whether you can think of additional analyses.
I scraped papers from PubMed published in JPR that mention a GitHub URL in the abstract (unfortunately it seems that even for OA papers the full text is not easily accessible) and calculated a few statistics. Sample size is 89 papers.
By definition unfortunately the majority of closed source software won't be included in this analysis because they won't list a GitHub URL in their papers.
Let me know if this is interesting and whether you can think of additional analyses.