Abstract:

In this paper we describe a visionary system, VAPD, which supports crime analysts in uncertainty aware decision making in use of comparative case analysis. In this scenario, it is crucial for crime analysts to get an accurate estimate of uncertainties included in their data as well as those caused through data transformations and mappings, thus supporting analysts in calibrating their trust in the pieces of evidence gained through data analytics. VAPD consists of one data processing and three visualisation components that adopt a set of guidelines for handling uncertainties. The system focuses on conveying an accurate estimate of these uncertainties on processes and uncertainties that occur within its natural language processing components. Text data analysis is ambiguous and error prone, but is nevertheless an important part of the data analysis. Through its innovative handling of uncertainties, VAPD enables

transparent and reliable decisions based on uncertainty-aware visual analytics.