Display Wall Empowered Visual Mining for CEOP Data Archive

Abstract

A large display space potentially improves data exploration, especially when where there are several types of data sets to control. To better understand the feasibility of a large display space, we implemented a display wall assisted visual mining system. A 5 m × 2.5 m display allows users to simultaneously visualize various types of data. The system also has a very high resolution of 15 XGA displays (each XGA display has 1024 pixel × 768 pixel resolution). With these attributes, the user can easily examine and visualize a dense data set in detail. To our knowledge, few systems of this type that are easy to use have been realized. A number of systems are now functional (e.g., the NASA AMES Hyperwall, http://www.nas.nasa.gov/Groupsl/VisTech/hyperwall/). However, these systems require much knowledge and power of the user. Most users of such systems have to use some special application, which can be difficult for earth environmental researchers. Although the proposed system has had limited testing by some researchers, we believe that our experience would be helpfu1 for users who are trying to develop similar systems. The proposed system is more efficient as it allows for simultaneous multi-dimensional visualization of complex data products, and various mining functions, such as two-dimensional correlation, are also prepared for the Coordinated Enhanced Observing Period (CEOP) data archive. By using these functions, various advanced analyses that are not easily utilized with conventional mining tools are attained. Moreover, several scenarios using the powerful capability of this tool are also presented. Finally, the process of data analysis in the proposed system becomes much more efficient through the use of a large-scale high-resolution display wall.