Designing Data-Driven Solutions to Societal Problems: Challenges and Approaches

Janardan Misra, Divya Rawat, Sanjay Podder

Abstract: Deciding effective and timely preventive measures to address societal problems is a difficult challenge. Societal problems tend to be inherently complex, affect the bottom of the socio-economic pyramid, and have wider spread across populations and geographies with resource constraints. In this paper we identify challenges and propose certain design considerations to guide development of machine learning based data-driven applications for addressing societal problems. Challenges cover subtle difficulties, which may be encountered during various phases of design and deployment life-cycles and design considerations make recommendations with respect to currently known state-of-the-art concepts, tools, and techniques to address these challenges.