Towards Personalized Instruction: Co-designing a Teacher-Centered Dashboard for Learning Engagement Analysis in Blended Learning Environments

Learning engagement is an important factor in academic success and instructional quality. In authentic settings, analyzing learner engagement status can be challenging due to the complexity of its multidimensional construct. It requires constant monitoring of multimodal engagement indicators (cognitive-level indicators, e.g., effort, emotional-level indicators, e.g., positive/negative emotions) and individualized pacing of learners, identifying when and who needs assistance, as well as making pedagogical strategy that matches the needs of the learners (or group of learners). This study combines user requirements with computational techniques to construct a multimodal learning engagement analysis framework in the blended setting. In addition, a teacher-faced dashboard prototype is developed to statistically summarize and visualize multimodal indicators in a way that enables teachers to deploy personalized instruction schemes. The teachers’ perspectives discussed in this study portray the great potential of introducing Artificial Intelligent (AI)-augmented models and visual analytics techniques aimed at deploying personalized instructions in the blended learning environment.