Leveraging visualization technology in Geospatial applications

As distinct technologies, geographic information systems (including geospatial data models), and visualization systems offer complimentary services and data management in application design. By combining them both into an application architecture, full feature sets and benefits of both approaches are realized. This paper introduces the relationship and differences between geospatial models and visualization technology. A series of short case studies present fully developed application scenarios where both geospatial and visualization technologies are well integrated. The case studies cover applications by an Electric Utility Company, three European Telecommunications Companies, and a North American Defense Contractor. As further discussion on how the technologies interact at a technical level, the paper discusses an interface between the Oracle Spatial Cartidge (multi-dimensional spatial database system) and the AVS/Express visualization software system. To conclude, the paper discusses rapid prototyping methodologies for integrating geospatial and visualization software subsystems.

Relationship of Geospatial and Visualization Technologies
It is important at the outset to define a frame of reference for geospatial and visualization technologies before we can discuss how they are integrated and used within an application. In traditional GIS applications, data is held in flat or proprietary structured storage systems. The data held in GIS repositories consists of three main types: raster or gridded data, vector elements for positional, shape, or coverage data, and multi-variate attribute data associated with geographically located positions. GIS applications typically process a large number of logical layers for a given area, often including data sources at different resolutions or scales. The display and review of GIS data is typically through direct graphics display to the window system, showing a subset of available 2D layers. The third dimension, when used, is generated by assigning an attribute such as elevation to create a 3D surface display.

The area of visualization technology focuses on the issues of processing a wide variety of data sources for interactive graphics display. Data models used for visualization are both spatial coordinate based, and nonspatial abstract data model based. The data models in visualization systems are generalized and abstracted for use in a wide range of fields. Visualization systems include many methods for creating abstract visualization objects suitable for graphics display. These are abstract representations of data, rather than direct display of data. This abstraction layer allows processing during visualization, such as contouring, interpolation, feature identification, and sampling from continuous fields [Wiss 98].

The abstract visualization object approach allows “Closed Loop” interaction. Rather than viewing visualization as a one-way process, direct mouse interaction with the graphics is what is used to drive further processing or changes of parameters of the data processing. The concept of using the visualization itself as the interactive user interface of the application is called “Visual User Interface” (VUI). It enables much more intuitive exploration of data than using secondary menu systems to steer the graphics display [Jern 99, Brown 95].

When geospatial and visualization systems are coupled together, each perform in areas where they are best designed. The geospatial operators manage data access, navigation, coordinate systems, and spatial analysis. The visualization system creates a rich graphics environment for representative visual treatment of data, and provides an interactive path for parameters and display refinement.

Electric Power Utility Application
An Electric Utility company in a major city is designing a specialized decision support application, coupling geospatial data models with advanced data visualization. The dual-aspect data source, interactive review, and the need to animate time history data generated requirements for a hybrid application design.

As the utility sector moves to a private, competitive marketplace, pressure is increased to effectively manage asset ROI and profit margins. Whereas Electric Utilities have traditionally managed power grids for long-life, reliability, redundancy in the case of failures, and over-designed duty cycles, the days of such long-term capital investment strategies are over.

Within the city, individual power transformers supply electricity to several blocks of residential & business customers. Typically each transformer is managed so that the power loading is at a fraction of its peak rating. In this way the life of the transformer is extended, and extra capacity is available for peak or emergency conditions. The closer the average load gets to the peak transformer rating, the shorter the equipment life cycle becomes. Striking the right balance between equipment cost, maintenance, and peak capacity is critical to obtaining a competitive supplier position.

The visualization application is designed specifically for this asset management and optimization problem. Historic transformer load data is accessed form a relational database. Transformer location, city map data, and spatial power consumption contours are derived from GIS sources. By choosing a near-peak day, such as when high temperatures indicate high airconditioning usage, hour-by-hour load factor data can be seen in an animated geospatial view. Key problem areas can be easily identified by the visual representation. By clicking on an individual transformer or an aggregated region, detailed equipment information is displayed, using “drill down” techniques.

The Electric Utility Company is able to plan and manage transformer upgrade and replacement programs much more effectively with a visualization-based decision support tool, giving managers a clear overview of critical load history data in a geo-located context.