Optimizing Virtual Learning for Classical Poetry: Insights From Spatial Analytics of Student Engagement

Abstract

Immersive virtual learning environments (VLEs) hold transformative potential for humanities education, particularly in subjects like classical poetry, which can be abstract for novice learners. To understand student engagement and optimize VLE design for fostering interest in Haiku, a traditional Japanese poetry form, it's crucial to move beyond traditional assessment methods. This study explores spatial behavior analysis as a means to evaluate student interaction within a custom VLE centered on a Japanese Haiku master's historical journey. We developed two types of VLEs featuring either locations tied to a recreation of the poet's journey embedded in a geographic 3D model. By tracking the position and orientation data of 208 informatics students as they navigated these spaces, we sought to determine if these metrics correlate with levels of engagement and how best to optimize VLEs for enhanced student interest. K-means clustering segmented students by interaction time with key elements, uncovering diverse exploration patterns. These ranged from targeted assignment completion to extensive engagement with supplemental content. Specific clusters also suggest areas where enhanced instructions may be needed. These findings empirically demonstrate how this approach enables educators to iteratively tailor VLE designs to individual learner needs and create engaging, immersive experiences that support the complex study of classical poetry. This study provides a model for leveraging spatial analytics to guide development and refinement of innovative learning experiences in virtual spaces.