Interacting with the Data Warehouse: Applying Virtual Reality Technology for Exploratory Data Analysis

Vaughn Stewart
School of Geography, University of Leeds, Leeds LS2 9JT, United Kingdom

Data analysis is completed by both industry and academia for studying trends or patterns in information. Industry refers to large corporate data stores as Data Warehouses. The two dimensional presentation features available in current Data Warehousing software interfaces limits the amount of data displayable and also the methods of user interaction. As Researchers and corporate data analysts study more sophisticated data phenomena they will expect and require more intuitive methods of interacting with their data.

Virtual Reality Technology (VRT) can be harnessed to create interactive graphical interfaces which can support four dimensional (3D and time) analysis. Although VRT has been forever embedded into the minds of the populace as technology for computer games and cyber punks it is not all head mounted displays and powergloves. VRT can be used to create Interactive Visualisation applications for the standard Windows (Microsoft, Motif) desktop requiring no special hardware. The Pentium personal computers available today have more than enough power to generate complex artificial worlds or Virtual Environments (VE). Using VRT powerful metaphors can be developed to model information in a visually expressive manner. Additionally, proper exploitation of the interface enables more characteristics to be displayed for a given information object, and more objects to be viewed.

This paper introduces the components of VRT (e.g. universe, sensors, objects), the features (e.g. behavioural modelling, direct data manipulation) in VRT which can be exploited to perform effective data analysis, and describes a prototype VE data analysis system which was developed to explore a retail Data Warehouse.

Knowledge of VRT components enables interfaces to be described with a standard set of terminology. Understanding VRT components and features will help designers create effective visualisations, and ensure users understand interface controls.

The prototype VEs developed by the researcher were programmed in C++ using the Sense8 VR toolkit for the Windows 95 operating system. Information was presented in the Data Warehouse visualisation system using stunning three dimensional graphic primitives. The user is free to touch, turn, and fly around to explore data objects from any viewpoint. Running simulations, the user can animate historical data to visualise trends developing over time and analyse the events which are driving the data patterns. Displaying information in this intuitive manner enables users to adapt and train their visual recognition skills and perform decision support.

The paper concludes with a discussion on the VRT industry and application areas in the retail market where VRT could play a major role in revolutionising methods in data analysis.