Diversity of Skills and Collective Intelligence in GitHub

Dorota Celinska-Kopczynska

Abstract: A common assumption suggests that individuals tend to work with others who are similar to them. However, studies on team working and ability of the group to solve complex problems highlight that diversity plays a critical role during collaboration, allowing for the diffusion of information. In this paper, we investigate the patterns behind the connections among GitHub users in Open Source communities. To this end, we use Social Network Analysis and Self-Organizing Maps as the similarity measure. Analysis of textual artifacts reveals the roles of those connections. We find that diversity of skills plays an essential role in the creation of links among users who exchange information (e.g., in issues, comments, and following networks). The connections in networks related to actual coding are established among users with similar characteristics. Users who differ from the owner of the repository report bugs, problems and ask for help more often than the similar ones.