Increasing categories of electronic communication platforms, such as instant messaging, Email, forums, and the like, have facilitated communication and collaboration among people worldwide. A large amount of online communication data generated by these platforms is collected and has accumulated, providing opportunities for analysts to understand communication patterns and facilitate decision making. An example is that, MOOC forums are becoming central hubs where students are able to interact with instructional staff. The analysis of MOOC forum data is beneficial to understanding class dynamics and preparing courses for the next iteration. However, such analysis is challenging due to the large, complicated, and heterogeneous nature of online communication data. Information visualiz...[ Read more ]

Increasing categories of electronic communication platforms, such as instant messaging, Email, forums, and the like, have facilitated communication and collaboration among people worldwide. A large amount of online communication data generated by these platforms is collected and has accumulated, providing opportunities for analysts to understand communication patterns and facilitate decision making. An example is that, MOOC forums are becoming central hubs where students are able to interact with instructional staff. The analysis of MOOC forum data is beneficial to understanding class dynamics and preparing courses for the next iteration. However, such analysis is challenging due to the large, complicated, and heterogeneous nature of online communication data. Information visualization has been proven effective in understanding enormous amounts of such data by turning it into visual representations to exploit the pattern recognition capabilities of the human visual system.

In this thesis, we propose three advanced visual analytics systems for understanding online communication data in various domains. The first system, iForum, is designed to investigate the three interleaving aspects of MOOC forums, that is, users, posts, and threads, at different granularities. Second, we present visForum, a novel visual analysis system for interactively exploring, comparing, and tracking conversation groups in online forums. The third system that we propose is T-Cal, an interactive visualization system to understand team conversation data from different perspectives. All three systems are able to provide insights into user communication behavior for analysts. To validate the effectiveness and usefulness of proposed systems, we conducted case studies involving domain experts for all three systems and one user study for visForum.