Plenary 2         GIS Representation for Visualizing and Mining Geographic Dynamics
Dr. May Yuan
http://geography.ou.edu/people/myuan.html
myuan@ou.edu
405-325-4293 (voice)

Geographic dynamics refers to the development of processes that shape our environment. While geographic dynamics is central to geographic understanding, the current GIS technology is incapable of capturing information that reflects the working of geographic processes, such as spread of a wildfire, development of a weather system, and propagation of a disease. Failure to represent geographic dynamics excludes GIS capability to mine information about spatiotemporal behaviors which can lead to new insights into understanding evolution and influences of geographic processes. Since geographic dynamics is, by definition, spatial and temporal, GIS data models must be able to handle temporal information to represent geographic dynamics. The current GIS data models organize data based on how data are collected and are location-based because geographic data are collected at a location or from an area. Geographic semantics (meanings) of a location are considered as attributes of the location. When time becomes necessary, geographic semantics of the location will be updated accordingly.

However, geographic dynamics requires an alternative approach to represent geography. One popular approach to discern geographic dynamics is through visualization: animate snapshot views of the process in a time sequence. In an animated view, the user identifies an object of interest and keeps track of its development. To automate identification of geographic dynamics in GIS, the alternative approach shall simulate the visualization process to represent geography. The focus should be placed upon geographic semantics that corresponds to processes. GIS data models should maintain spatial properties and attributes of these processes over time, and hence represent geographic dynamics. In doing so, GIS can provide direct support to analyze and visualize spatiotemporal behaviors of processes, retrieve and compare geographic processes, understand the dynamics involved in geography. The presentation will outline the alternative approach in contrast with other spatiotemporal GIS data models, examine hierarchies of geographic semantics, space, and time, and suggest a research agenda to advance geographic representation that can facilitate visualization and mining geographic dynamics.