Perceptions of project-based learning in CDIO-aligned curricula: an exploratory cluster analysis of a five-university survey in South Africa

Abstract

Engineering education in Africa is undergoing a transformative shift, seeking to align with global standards while addressing the unique challenges and opportunities of the region. The integration of innovative teaching methodologies, such as Project-Based Learning (PBL), has emerged as a promising strategy to bridge the gap between theoretical knowledge and practical application. Within this paradigm, the CDIO (Conceive, Design, Implement, Operate) framework provides a structured approach to developing engineering competencies that are both technically robust and contextually relevant. However, evaluating the effectiveness of such educational strategies in diverse and resource-constrained environments remains a significant challenge. Unsupervised machine learning offers a powerful and scalable solution to this challenge. By analyzing complex, multidimensional data without predefined labels, it can uncover hidden patterns and insights into how students engage with PBL methodologies under the CDIO framework. This approach enables a nuanced understanding of student and staff behaviors, learning outcomes, and contextual factors that influence success through capacity building.   This article explores how an unsupervised machine learning approach can be applied to understand and enhance the implementation of PBL using the CDIO framework in African engineering education. Adapting these data-driven insights from a five-university South African sample, this study proposes actionable recommendations for educators and policymakers, while noting that broader generalisation beyond similar contexts requires further multi-country validation