Software analytics study of Open-Source system survivability through social contagion

Free, Libre and Open Source Software (FLOSS) development has grown in prominence in recent years as the software development approach of choice. However, the factors responsible for maintaining development interest in them are still poorly understood, as FLOSS projects differ significantly from traditional for-profit closed-source software development. To address that knowledge gap, we have analyzed data from GitHub, an open source code repository which provides extensive records on a multitude of collaborative software engineering projects. Using machine learning algorithms, we sought out patterns in the data that might help us understand how projects survive. Our findings suggest that the impressions propagated by the users of software projects to other users strongly influence the projects´ survival, as one would expect in a social contagion model.