Towards data-driven interventions: Leveraging learning analytics to support programming education for grade 10 learners in South African schools

Abstract

Programming is increasingly incorporated into school curricula worldwide to foster essential 21st century skills.
However, many educational systems face challenges in integrating it effectively because of limited resources
and support. In South Africa, a lack of tools further compounds these challenges, making it difficult for teachers
to identify and address learners’ specific needs. In recognising these challenges, we aimed in this study to
develop and validate a Learning Analytics (LA) model to identify challenging programming concepts for Grade
10 learners in South Africa. Using the LA five-step model, we employed Microsoft Power BI for its analytical,
visualisation, and AI-driven forecasting capabilities to analyse historical examination data systematically. The
resulting forecasting model identified five key areas in which learners struggle: conditional statements; problem
conceptualisation/solution design; debugging/exception handling; abstraction/pattern recognition; and
class/object differentiation. Our findings demonstrate the potential of LA-powered models to guide targeted,
data-driven interventions, supporting improved learning outcomes, and engagement in programming for Grade
10 learners.