(MESSY) BIG DATA FOR SOCIAL GOOD

The goal of this work is to develop algorithmic approaches that tackle complex social problems. All three projects included in the dissertation start with large-scale messy datasets in social domains with sensitive nuances that are important for considering when designing algorithmic analyses. Our primary focus areas are human rights abuses in supply chains andcounter-sex trafficking investigations. We combine datasets from dozens of sources to provide unique views into these domains. We develop algorithmic methodologies that account for the distinct challenges for each area and share how these contributions can be applied to other domains.