Analysing Associations of Textual and Relational Data with a Multiple Views System

Information access is developing from data- to content-oriented access, as the hype of semantic web technologies confirms. Combining the access to textual and database data is one aspect on this way. In the paper at hand we address the explorative analysis of text documents that are associated with relational tuples, which is particularly important for application fields like customer relationship management or business intelligence. We present a system for comprehensive visual analysis based on a multiple views paradigm which brings together views on text similarity, text categories, and associated relational attributes. For keeping track of navigation and selection operations in complementary views, we introduce the application of visual overlay techniques. Moreover, we describe a tree compression technique for allowing multiple focus selections on text classification trees. We sketch an application example, discuss the solution with respect to design guidelines from literature, and describe results of a usability evaluation.