A Big Data framework for analysing educational data to enhance student retention at South African universities

Abstract

Background: Universities in South Africa continue to face persistent challenges in improving student retention. Existing support interventions are often fragmented and reactive, lacking timely mechanisms to identify and assist at-risk students.
Objectives: This study proposes a Big Data framework to support the analysis of educational data for early risk identification and targeted intervention, thereby enhancing student retention in South African universities.
Method: The study adopted an interpretivist qualitative approach, supported by Design Science, integrating qualitative data from interviews with 30 academic staff members across four public universities, along with the analysis of existing institutional retention reports, policy documents, and strategic plans documentation.
Results: Findings of this study revealed fragmented data systems, limited data literacy among staff, and concerns regarding the ethical use of data. Participants expressed the need for predictive analytics, integrated data systems, and personalised support mechanisms as key enablers of improved retention strategies.
Conclusion: The study concludes that Big Data analytics can play a critical role in improving student retention rates in public higher education institutions in South Africa. The findings demonstrate that when institutions integrate diverse student datasets, apply predictive modelling, and use real-time monitoring tools, they are better positioned to identify students at risk of dropping out and implement timely, evidence-based interventions.
Contribution: The proposed framework enables institutions to shift from reactive to proactive student retention management through ethical data governance, integrated data processing, predictive analytics, and personalised interventions.