Analyzing developer behavior and community structure in software crowdsourcing

Abstract

Recently software crowdsourcing has become an emerging development paradigm in software ecosystems. At present, most research efforts on software crowdsourcing focus on modelling its competitive nature from aspect of incentive mechanism and developer decision using game theory. Few work has been done on analyzing the impact of developer behavior and community structure on software crowdsourcing practices. Based on social network modeling and analysis methodology, this paper studies a popular community of software crowdsourcing named as TopCoder. We discover that the online activities of TopCoder users can be characterized as a temporal bursty pattern. Such a user behavior leads to the similarity in the participants of TopCoder contests occurred within the consecutive time intervals. Furthermore we introduce a competition social network to study the influence of the cooperation and competition between developers on their rating in the TopCoder community by analyzing competition social network. In addition to the community-wide developer network, this paper extends the modeling method to examine topologic characteristics of TopCoder projects and reveal the correlation between these structural features and the outcome quality of the projects.