Hidden in the Code: Visualizing True Developer Identities

Analyzing developers' activity is an essential step in understanding a software project. When assessing the status of a project and studying its history and development, it is vital to assess the performance of each contributor with high precision. Unfortunately, as Version Control Systems (VCS) do not manage the concept of developer identity in a precise fashion, it is often hard to connect a single identity with all the contributions a user has made to the project, by taking into account potential aliases of the same person. Our work aims to bridge this gap by processing the information related to the identities of contributors in the most popular projects available on GitHub, and to visualize their activity over time. The visualized identities are built on top of a disambiguation algorithm that aims to merge aliases of the same person into a single entity. Moreover, we perform bot detection to differentiate the activity made by bots from the one of humans. At last, we cluster the resulting identities to find users with the same activity patterns. By interacting with the proposed visualizations, one can precisely analyze the contributions of the people working on a project. Video URL: https://youtu.be/098IsBDBXKY