Sunbelt 2014: Temporal Visualization of Dynamic Collaboration Graphs of OSS Software Forks

In this work, we studied collaboration network of three open source projects using a combined analysis method of temporal visualization and temporal quantitative analysis. We based our study on two papers by [Robles and Gonzalez-Barahona 2012] and [Hanneman and Klamma 2013], and identified three projects that had forked in the recent past. We mined the collaboration data, formed dynamic collaboration graphs, and measured social network analysis metrics over an 18-month period time window. We also visualized the dynamic graph (available online) and as stacked area charts over time. The visualizations and the quantitative results showed the differences among the projects in the three forking reasons of personal differences among the developer teams, technical differences (addition of new functionality) and more community-driven development. The personal differences representative project was identifiable, and so was the date it forked, with a month accuracy. The novelty of the approach was in applying the temporal analysis rather than static analysis, and in the temporal visualization of community structure. We showed that this approach shed light on the structure of these projects and reveal information that cannot be seen otherwise.

Print Friendly, PDF & Email
This entry was posted in All, Publications and tagged , , , , . Bookmark the permalink.

Leave a Reply