Exploring Large Geo-Spatial Point Sets

possible
because of rapid advances in the technology of processors, storage and networks. As a result, database management systems store
and
exchange extraordinarily vast data sets, of terabytes and beyond. The analysis of patterns and relationships in geo-spatial
data is a
central function of geographical information systems (GIS). The difficulty lies in finding the details that reveal fine
structures
hidden in such vast data sets. Displaying large point sets on conventional maps is problematic. Overplotting obscures data
points in
densely populated areas; however, sparsely populated areas waste space while conveying scant detailed information. To cope with
this
scale, we are interested in combining visual and algorithmic data mining for the exploration of large point sets on maps.

In this talk, we will introduce the PixelMap problem and a efficient heuristic (called Fast-PixelMap) for its solution. In
addition, this talk presents some latest research results in generic data layouts for large geo-spatial points sets, and
finally some
challenges and research problems in exploring space-time pattern will be discussed.