Representing, analyzing, modeling and extracting meaning from complex heterogeneous geospatial datasets requires new approaches that can scale up to current and future data complexity and data volume. Our work addresses a wide variety of issues, including:
- the development of 'complex' spatiotemporal systems with emergent properties,
- new techniques for data mining, knowledge discovery, visualization (for application to geospatial and spatiotemporal information about the past, present, and future),
- advanced and semantically aware spatial databases that can represent and integrate both the data and the various higher level knowledge constructs, such as categories and relationships that emerge from the data during knowledge construction and
- developing a geographical agent modeling environment for investigating human activities.
These activities, when integrated, support the entire geo-scientific process, from initial exploration of data, hypothesis generation, concept discovery, model formulation, analysis and validation, and, when fused together seamlessly in GeoVISTA Studio, will form a complete Problem Solving Environment (PSE) for teams of scientists to use, thus supporting our geocollaboration focus. By bringing these activities together in GeoVISTA Studio we avoid many of the integration problems that plague traditional computational analysis. To accomplish this goal, Center affiliates and their collaborators are working to integrate methods and tools that span many disciplines including machine learning, pattern recognition, agent and cellular modeling, data mining, multivariate information visualization and spatial statistics.