Understanding Learner Interaction Analytics through Data Visualization in Metaverse-Based Learning Environments

Abstract

Metaverse-Based Learning Environments (MBLEs) generate rich, high-volume multimodal interaction data (e.g., gaze, gesture, speech, movement) that can reveal learner engagement patterns. Yet, instructors lack tools that can cluster and visualize these data in an actionable, real-time manner. This doctoral research aims to develop a data visualization framework that integrates unsupervised clustering of multimodal student interactions with adaptive, immersive dashboards to support pedagogical decision-making. The study proceeds in three phases: (1) establishing a validated operational framework for MBLEs, (2) designing and evaluating a clustering pipeline for multimodal learner data, and (3) prototyping immersive dashboards that reduce cognitive load and enhance instructional responsiveness. Preliminary outcomes include a Unity-based data capture prototype. This work advances immersive analytics, multimodal learning analytics, and Human–Computer Interaction by linking embodied learner behaviors to adaptive visual feedback.