Our project proposal seeks to develop an analytical system that optimizes course design based on real-time student learning behavior data. By integrating educational theory frameworks, the project aims to utilize student learning data effectively and assist teachers in dynamically optimizing learning tasks and course modules. The key objectives are three-fold: 1. Enhancing Learning Outcomes: Analyzing student behavior data to adjust course content in real-time for personalized and efficient teaching. 2. Implementing Advanced Educational Theories: Integrating cutting-edge educational theories with data analysis for scientific course design. 3. Visual Analytics and AI Integration: Using visual analytics and AI algorithms for deep analysis of student learning data, offering real-time feedback and optimization suggestions.The project aims to equip educators with an effective tool to leverage real-time data in course design, catering to individual student needs and enhancing overall educational quality.

Our project proposal seeks to enhance personalized education by using generative AI, like GPT-4, to analyze student interactions and identify knowledge gaps. By understanding these gaps, teachers can design tailored learning plans and exercises. A major challenge is managing the vast, unstructured chat data, as well as enabling effective collaboration between teachers and AI to create diverse exercises. Visual analytics will be employed to simplify data interpretation, allowing teachers to grasp student progress and efficiently verify AI-generated exercises. The research will concentrate on university-level data visualization courses, highlighting complex, multidisciplinary knowledge structures. The project will proceed in three stages: developing visual analytics tools, generating personalized exercises, and evaluating the system in real-world scenarios. This approach aims to improve educational quality, meet individual student needs, and explore new methods for integrating AI into teaching strategies.

Our project proposal seeks to revolutionize English language education by addressing common challenges faced by learners. By bringing together the expertise of academic institutions, technology companies, and schools, our project aims to empower English language learners, offering them an immersive language tutoring system within the metaverse. By leveraging AI technology, an AI tutor will provide learners with prompt and expert feedback, allowing them to enhance their proficiency in real-time. Implemented within a virtual reality environment, the system will enable learners to practice and refine their English communication skills through simulated real-world scenarios. This innovative approach aims to create an engaging and authentic learning experience, enhancing language proficiency and building learners' confidence in using English in practice.

Teaching computational thinking methods is vital in preparing K-12 students ready for the information age. Online learning platforms supplement traditional schooling but lack efficient monitoring and guidance due to information gaps and time constraints. Therefore, we proposed to revolutionize K-12 education and empower students to excel in the technology-driven landscape of the information age by reducing the workload of teachers and parents, facilitating efficient analysis of student performance, and providing personalized guidance. To achieve this goal, we will design a visual analytics system; innovate an automated reporting system; and invent a personalized learning path recommendation system. Our work will be deployed and rigorously tested in real-world settings, ensuring tangible and impactful results.

Virtual communication continues to demonstrate its significance in a post-COVID world. However, measuring and analyzing audience engagement in this context remains challenging. Our project aims to develop an analytics method for assessing audience engagement in virtual communication. Through three modules - feature extraction, data labeling, and engagement analytics - we collect multi-modal data. By merging these features and utilizing machine learning models, we extract valuable insights into audience engagement. This information benefits both speakers and listeners by facilitating the evaluation of presentation effectiveness and understanding of communication behaviors. The techniques will be applied and evaluated on popular call-tech platforms such as Zoom, Microsoft Teams, and Google Meet, potentially impacting a range of industries.

This project developed visualization techniques to analyze large-scale data from MOOCs, assisting course instructors and education analysts. The techniques allow exploration of various aspects, including video popularity, clickstream analysis, demographic distribution, user forum interactions, dropout analysis, and sequence analysis. An innovative interaction data collection technique was created to collect fine-grained mouse movement data during problem-solving on a K-12 E-Learning platform. The data was used to analyze problem-solving sequences and develop intuitive visualization systems like QLens and SeqDynamics, helping infer students' cognitive states. The main goal of this project is to provide intuitive and interactive analysis, allowing stakeholders to understand learners' behaviors and improve course designs with evidence-based guidance.