Generating developer portfolio based on coding behaviour using GitHub mining and machine learning

Abstract

The widespread use of open-source software repositories, such as Github, provides a rich data source for analyzing developer coding behavior. Manually curating developer portfolios to analyze their skills, contributions, and coding patterns would take a lot of time and be challenging. So, we will create a framework to automatically generate developer portfolios by mining Github repositories and applying machine learning techniques to model classify developer skills and predict work patterns.. The framework makes use of Github API data like commit history, pull requests, metrics of code complexity and collaboration patterns to extract features that help identify the developer coding styles and abilities. Machine learning algorithms, including clustering and classification, will use their technical strengths, domain background and quality of code to segment developers. The goal of producing these portfolios is to assist recruiters, project managers, and developers in better understanding the skill sets that exist within the network. As well as improve career growth strategies and identify opportunities for collaboration.