As problem-solving skills are becoming increasingly important for citizens, many question-based learning platforms emerge to provide students with online questions to cultivate their problem-solving skills. For example, various MOOC platforms (e.g., Khan Academy), online question pools (e.g., LeetCode) offer interactive maths questions and programming exercises. The distance gap, time gap, and the imbalanced number of educators and students make it hard for educators and students to communicate in the problem-solving processes. For educators, they need to understand students’ problem-solving logic and learning habits to improve question designs and give customized instructions to groups with different cognitive abilities and non-cognitive traits. For students, they can improve...[ Read more ]

As problem-solving skills are becoming increasingly important for citizens, many question-based learning platforms emerge to provide students with online questions to cultivate their problem-solving skills. For example, various MOOC platforms (e.g., Khan Academy), online question pools (e.g., LeetCode) offer interactive maths questions and programming exercises. The distance gap, time gap, and the imbalanced number of educators and students make it hard for educators and students to communicate in the problem-solving processes. For educators, they need to understand students’ problem-solving logic and learning habits to improve question designs and give customized instructions to groups with different cognitive abilities and non-cognitive traits. For students, they can improve their self-learning skills based peers’ learning data. For example, regulating their learning habits and planning personalized learning paths. However, it is challenging for educators and students to analyze and learn from thousands of problem-solving sequences. Visualization technologies turn out to be an effective solution to support data presentation and analytics in the aforementioned scenarios.

In this thesis, we enhance the communication between educators and students in the context of the online problem solving by a visualization approach. On the one hand, we present two visual analytics systems for educators to understand students’ problem-solving behaviors from two levels respectively. The first system, QLens, helps question designers analyze students’ problem-solving behaviors in multi-step questions to improve question designs at a micro-level. The second system, SeqDynamics, evaluates students’ problem-solving dynamics from both cognitive and non-cognitive perspectives at a macro level. On the other hand, we try to improve their self-learning skills, which includes learning regulation and learning planning. We use information visualization to promote students’ reflection on “Gaming the system” behavior and reduce the gaming behaviors. In addition, we propose PeerLens, an interactive visual analytics system that enables peer-inspired learning path planning. We have conducted various quantitative evaluations, case studies, user studies, and expert interviews to demonstrate the effectiveness and usefulness of our proposed visualization approaches.