A large-scale analysis of bioinformatics code on GitHub
Fig 2
Project features by article topic.
Projects are broken into groups according to whether the accompanying paper abstract is associated with each topic category. Projects that are associated with multiple topics are counted separately for each topic. Topic labels were assigned manually after examining top terms associated with each category. We added one to several variables to facilitate plotting on a log scale; these are noted in the variable name. All variables refer to the GitHub repository except “1 + mean PMC citations / week”, which refers to the paper and looks at citations in PubMed Central per week starting two years after the initial publication of the paper. Commits is the total number of commits to the default branch. Commit authors have created commits but do not necessarily have push access to the main branch; we attempted to collapse individuals with multiple aliases. Forks are individual copies of the repository made by community members. Subscribers are users who have chosen to receive notifications about repository activity. Stargazers are users who have bookmarked the repository as interesting. Megabytes of code and total files include source code only, excluding data file types such as JSON and HTML. The horizontal line at the center of the notch corresponds to the median. The lower and upper limits of the colored box correspond to the first and third quartiles. The whiskers extend beyond the hinges by at most an additional 1.5 times the inter-quartile range. Outliers are plotted individually. The notches correspond to roughly a 95% confidence interval for comparing medians [27]. The table of repository features is provided as S8 Table.