JExpert

Abstract

Software development, maintenance, and evolution are increasingly challenging tasks and require skilled professionals in many different technologies. Enterprise and open source projects seek to build the best possible team composed of highly skilled developers in specific libraries. On the other hand, the identification of such skilled professionals is not a trivial task. In this paper, we introduce JExpert, an automated tool that identifies library experts from source code. This tool is designed to identify experts in specific libraries based on source code activities from GitHub projects. In a preliminary evaluation, we rely on JExpert to identify the top experts in 6 libraries that support microservice-based application development. In total, we analyzed more than 1,200 projects and 797 developers. JExpert outputs a summarized profile of each expert with 3 expertise metrics, namely, number of commits, imports, and lines of code written for the library.