| A spatial data warehouse (SDW) is defined as a subject-oriented, integrated, time-variant, and non-volatile collection of both spatial and non-spatial data in support of the management's decision-making process. In this dissertation, we study two major research issues facing the SDW, including (1) a novel data warehouse model with associated data analysis operations, (2) selective materialization cost and increase the query efficiency.; First, SDW poses many challenges with respect to the data model and the processing of analytical operations. This is primarily due to the multi-dimensionality nature of each dimension that comprises of both spatial and non-spatial attributes. Moreover, the traditional OLAP operations are not adequate to specify and process the complex analysis queries posed to an SDW. Therefore, we present a suitable model, cascaded star model, which is an extension of the star schema. We also propose the cascaded OLAP (COLAP) operations that enable specification of analysis queries on the cascaded star model including cascaded-roll-up, cascaded-drill-dawn, cascaded-slice, cascaded-dice and mcube.; Second, we address the issues of the optimal selection of views to be materialized in SDW. The spatial data is typically larger in size, which leads to high maintenance cost, and the spatial operations are more expensive to process. We develop graph based query transformation rules by embedding the cost value of each spatial operation into the query graph, and present a greedy algorithm for materialized view selection for a given SDW so at the local cost optimality can be achieved. Furthermore, we investigate the utilization of spatial metadata to further optimize the materialization cost of an SDW. We propose a new notion, preview, for which both the materialization and on-the-fly costs are significantly smaller than those of the traditional views. Essentially, a preview pre-processes the non-spatial part of the query, maintains pointers to the spatial data, and exploits the hierarchical relationships among the different views by maintaining a universal composite lattice.; We implemented an SDW prototype on cascaded star model and cascaded (SLAP operations by using real world environment data sets from Meadowlands Environment Research Institute in New Jersey. |
