Visual Search and Analysis in Complex Information Spaces - Approaches and Research Challenges

One of the central motivations for visual analytics research is the so-called information overload - implying the challenge for human users in understanding and making decisions in presence of too much information 37. Visual-interactive systems, integrated with automatic data analysis techniques, can help in making use of such large data sets 35. Visual Analytics solutions not only need to cope with data volumes that are large on the nominal scale, but also with data that show high complexity. Important characteristics of complex data are that the data items are difficult to compare in a meaningful way based on the raw data. Also, the data items may be composed of different base data types, giving rise to multiple analytical perspectives. Example data types include research data compound of several base data types, multimedia data composed of different media modalities, etc. In this paper, we discuss the role of data complexity for visual analysis and search, and identify implications for designing respective visual analytics applications. We first introduce a data complexity model, and present current example visual analysis approaches based on it, for a selected number of complex data types. We also outline important research challenges for visual search and analysis we deem important