It is the only type of spatial dimension supported by nonspatial OLAP.
This type of spatial dimension is treated like other descriptive dimensions causing
the spatiotemporal analysis to be potentially incomplete and the discovery of certain
spatial relations or correlations between the phenomena under study to be missed
by the analyst. The two other types of spatial dimensions aim at maximizing the
potential to discover spatial relations and correlations that do not fit in predefined
boundaries. The geometric spatial dimensions comprise, for all dimension members,
at all levels of detail, geometric shapes (e.g., polygons to represent city boundaries)
that are spatially referenced to allow their dimension members (e.g., New York) to be
visualized and queried on maps. The mixed spatial dimensions comprise geometric
shapes for a subset of the levels of details. The members of the geometric and mixed
spatial dimensions can be displayed on maps using visual variables that relate to the
values of the different measures contained in the datacube being analyzed. Figure
3 presents examples of the three types of spatial dimensions.
Two types of spatial measures can be defined (B?©dard et al., 2001; Han et al.,
1998; Rivest et al., 2001; Stefanovik, 1997; Tchounikine et al., 2005). A first type
Figure 3. Three types of spatial dimensions: nongeometric, geometric, and mixed
spatial dimensions (Modeled on Rivest, B?©dard, Proulx, & Nadeau, 2003)
0 B?©dard, Rivest, & Proulx
Copyright ?© 2007, Idea Group Inc.
Pages:
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567