Presentation
The Student Retention Analytics Project represents a strategic shift from manual, reactive engagement monitoring to predictive, prioritised early intervention at scale.
Historically at the University of East London, over 23,000 students’ engagement data was reviewed manually during each teaching term, limiting the university’s ability to target support proportionately. The Student Retention Analytics Project consolidates attendance, virtual learning, environment activity, assessment engagement and student support data within a central Data Lake and applies pattern-matching to generate a withdrawal risk score (0–1). This score reflects how closely a student’s current engagement patterns align with those of previous students who withdrew.
Using a structured RAG framework and a defined threshold, formal low-attendance interventions have been prioritised from a cohort of approximately 5,000 students to fewer than 800 highest-risk cases, enabling more timely and targeted support while remaining policy compliant.
The session will explore the project’s development from Amazon Web Services (AWS) supported pilot to internally embedded institutional capability, including governance, ethical design (behaviour-based rather than demographic profiling), and how professional judgement remains central. It will offer practical insight into embedding predictive analytics within Student Services operations to support retention strategy and institutional performance targets.