Data Scientist University of Witwatersrand, United States
Session Abstract: Equity gaps in student persistence and completion remain a global challenge. This session explores how AI and predictive analytics can be designed to identify at-risk students early and support equity-focused interventions. Drawing from real-world implementations, we will demonstrate how integrating national assessments, socio-economic indicators, and institutional data produces actionable risk scores that power dashboards, nudging campaigns, and targeted advising. While case studies highlight South African contexts, the methods are broadly transferable to U.S. and international institutions. Attendees will gain practical strategies for building and validating equity-driven predictive models, insights into aligning these models with ethical governance frameworks (e.g., FERPA/POPIA), and of how data-informed nudges improve persistence outcomes. Participants will leave with replicable approaches to move from monitoring equity gaps to closing them through AI-enabled decision support.
