Análise de Redes de Colaboração em Comunidades de Desenvolvimento de Software: Um Estudo Baseado em Dados do GitHub

Abstract

With the continuous evolution of collaborative development platforms, new challenges arise in understanding how different forms of interaction shape the dynamics between developers and communities throughout the software development lifecycle. This study investigates the structure of collaboration within contemporary development communities, focusing on interactions recorded on GitHub. By collecting data through the GraphQL API and employing an ETL process to migrate information from MongoDB to Neo4j, a graph is constructed to integrate developers, artifacts, and contribution events. Network analysis algorithms—such as centrality measures, community detection, and structural metrics—reveal, for instance, that some activities are highly concentrated among a few contributors, that there are intermediaries between subgroups, and that dense communities form around Pull Requests. The results indicate that graph-based modeling offers an analytical perspective capable of uncovering technical and social patterns that characterize collaboration in distributed projects, thus contributing to broader investigations on organization and collective behavior within software development ecosystems.