Progressive Interactive Event Sequence Analytics: Combining Explorative Mining and Visual Reasoning

Temporal event data comprise sequences of point or interval events occurring over time and are today produced by a vast and growing number of data driven applications across both society and industry.

Effective analysis of this data can enable analysts to gain crucial understanding of complex and interconnected processes.

To this end, in this proposed project, we plan to perform innovative research at the intersection of temporal data mining and interactive visualization and produce visual analytics methods that facilitate progressive user-driven analysis of large and complex temporal event data sets.

More specifically, the research will focus on: (1) interactive, context-aware sequence mining methods for pattern identification in static and streaming event sequences, (2) visual reasoning approaches to communicate and compare how patterns appear and vary over time in event-sequence datasets, and to visually assess the causal relationships implied by the identified patterns, and (3) the design of interactive visual interfaces for communicating and interacting with the analysis process and the identified results.

The methods resulting from the proposed research programme will be integrated into a framework for testing and deployment on real-world data from selected application domains, including air traffic control, air traffic management, and analytical sociology.