An AI-Enhanced Learning Analytics System for Monitoring Multi-dimensional Student Engagement: Toward Data-Driven Smart Education

Student engagement is a central factor in academic success and the quality of learning experiences. Yet, most current approaches to monitoring engagement rely on static, single-point surveys that fail to capture its dynamic and multidimensional nature. This paper aims to explore the potential of data-driven systems in continuously monitoring and analysing student engagement, thereby enhancing both the learning experience and the learning environment. A cloud-based, AI-enabled engagement system and deployed for secure access by students and lecturers. Engagement data from 77 undergraduate students were collected over half a semester through structured self-reports and attendance monitoring. Data were analysed using descriptive statistics, Pearson correlation coefficients, t-tests, and longitudinal case studies. Results highlight strong emotional engagement, with Interest and Enjoyment showing the highest correlation (r = 0.798, p < 0.001). In contrast, Persistence & Effort correlated only weakly with Asking Questions (r = 0.263, p = 0.098). Box plots showed consistently high attention and effort but greater variability in peer interaction and question-asking. Longitudinal analysis revealed contrasting trajectories: one student demonstrated high, multidimensional engagement, while others repeatedly defaulted to mid-range scores. The identification of such passive “default” responses is a critical finding, highlighting potential disengagement and limitations of self-reporting when students do not actively reflect on their engagement. These findings demonstrate the value of continuous, technology-enabled engagement monitoring for capturing authentic learning behaviours. The system supports timely pedagogical interventions, early identification of at-risk students, and contributes to the development of intelligent, IoE-enabled educational technologies that promote sustainable learning outcomes.