Open Source Software for Social Good (OSS4SG) projects are projects that address a societal need and target people who need help. These projects often address high-impact humanitarian causes such as curating local health resources during a global pandemic, informing the public on the structural integrity of buildings, and encouraging civic engagement in times of strife. These projects carry a high intrinsic reward for contributing but are hard to find—prior research has shown that one of the the top challenges for contributors is not knowing where to find good projects to work on. Currently, contributors must manually search and assess whether projects align with their growing technical skills and intended impact interests.

In this paper, we describe a recommendation system that automatically recommends OSS4SG projects for contributors based on their activity and project-related information. To score and rank projects, we calculated scores based on four signals: technical skills, interests, social ties, and recency of project activity. We performed an offline validation of the recommendation system using standard evaluation metrics such as the hit rate ratio. Results show that the signals are effective in producing a ranked list of OSS4SG projects for contributors, with room for improvement. Finally, we conducted a formative study with contributors to better understand their process of project discovery, validate our findings, and identify additional signals for future work to improve recommendations.