University of Houston

Large spatial and spatio-temporal datasets with lots of variables pose challenging problems on how to store, query, summarize, visualize and mine such datasets. In this research we center on gridded spatial and spatio-temporal datasets which are quite common in scientific computing. One interesting and important task when analyzing such datasets is to find interesting contiguous regions, called interestingness hotspots, based on the domain expert’s notion of interestingness. In this research, we present a computational framework which uses a non-clustering approach to obtain interestingness hotspots. We present a novel hotspot growing algorithm which grows interestingness hotspots from seed hotspots. We claim that our approach is capable of identifying a much broader class of hotspots, which cannot be identified by traditional distance-based clustering algorithms.

Additionally, our research focuses on creating hotspot summaries, identifying characteristics of interestingness hotspots, and potentially on the visualization of high dimensional gridded datasets in general. Moreover, as datasets in this field tend to be quite large, parallel versions of the developed hotspot discovery and summarization framework will also be investigated.

We also investigate scoping algorithms for hotspots in 2D, 3D, and 4D. Scoping algorithms are particularly important for change analysis and efficient representation of spatial clusters and for hotspot summarization. In our future work we plan to generalize our work on two-dimensional scoping algorithms for 3D and 4D.