Requirement-Driven Developer Recommendation Framework Based on Github Developer Social Network

Developer recommendation is a field where focus is often made on the question and answer forums, bug tracking and resolutions and version control system commits. In this paper, a framework is proposed to utilize the huge information contained in GitHub repositories with respect to developer profiles and their activities. The proposed framework extracts and formalizes some of their features to create a Developer Network Graph (DNG). A novel ranking algorithm named Developer Experience and Requirement Attribute Algorithm (DERA) is proposed that uses developer experience and user requirements to rank and recommend developers. The proposed framework uses both feature weights of developers and their neighborhood scores to rank them, according to the requirements made by the user and display the result in real time. The model is experimentally validated on a dataset of 5000+ users and 200 languages. Results for different requirement scenarios are discussed for substantiation. The complexity of the proposed ranking algorithm (DERA) is identified as O(km) for k filtered users and m languages and is adaptable to real-life scenarios of finding developers that suit the business/project requirements.