Conference site ยป Proceedings

Testing Generative Models of Online Collaboration with BigBang

Sebastian Benthall
UC Berkeley School of Information

Abstract

We introduce BigBang, a new Python toolkit for analyzing online collaborative communities such as those that build open source software. Mailing lists serve as critical communications infrastructure for many communities, including several of the open source software development communities that build scientific Python packages. BigBang provides tools for analyzing mailing lists. As a demonstration, in this paper we test a generative model of network growth on collaborative communities. We derive social networks from archival mailing list history and test the Barabási-Alpert model against this data. We find the model does not fit the data, but that mailing list social networks share statistical regularities. This suggests room for a new generative model of network formation in the open collaborative setting.

Keywords

mailing lists, network analysis, assortativity, power law distributions, collaboration

Bibtex entry

Full text PDF