Effective computational research for social impact requires considering the entire pipeline from research to practice: project scoping, solution design, implementation, and evaluation. Ideally, the research will be carried out in partnership with practitioners and ultimately be used by the relevant stakeholders. However, the challenges of forming sustainable partnerships and the demands of academic publishing often preclude this work from achieving true benefit to society, especially for applications that lack strong pathways to profitability or primarily benefit marginalized communities.
We will discuss best practices and lessons learned from conducting research for social impact. In particular, we will highlight through three case studies how to build effective partnerships, identify impactful problems, and advance fundamental research while also moving towards deployment.