Abstract <p>As software projects increase in complexity, platforms like GitHub generate large volumes of data, including code commits, pull requests, and issue discussions. However, conventional tools often present these elements in isolation, limiting users’ ability to interpret development processes from multiple perspectives. We introduce RepoTale, an interactive visual analytics framework designed to support repository exploration by combining data integration, natural language processing, and visual analysis approaches. RepoTale is built upon a knowledge graph that links issues and pull requests, while applying topic modeling to identify thematic groupings within issue discussions. The framework features several coordinated visualization views to assist the investigation of project evolution and team interactions. We demonstrate RepoTale’s usability and effectiveness through case studies on real-world open-source projects, where the framework helps both newcomers and experienced contributors uncover interconnected patterns and solve practical problems efficiently. In addition, a comparative user study was conducted, showing that RepoTale improves task efficiency, insight generation, and user satisfaction compared with standard repository tools.</p> Graphic abstract <p></p>