Spatial OLAP Concepts

In order to support automated drill operations between the spatial levels of a spatial dimension, a particular configuration must be used to link the multidimensional data and the spatial data. By linking the spatial members of a datacube to the geometries of objects stored in a spatial database, these objects can be viewed and manipulated on maps. Such a structure results from the complete integration of the cartographic and the multidimensional components of a SOLAP solution and allows for interactive multidimensional exploration of phenomena.

There are various types of spatial dimensions and measures supported by Spatial OLAP tools and they are presented below.

The major part of the information of this page is from Rivest, S. et al. [5].


Spatial Dimensions:

A SOLAP system supports three types of spatial dimensions [1]: the non-geometric spatial dimensions, the geometric spatial dimensions and the mixed spatial dimensions. The non-geometric spatial dimensions use nominal spatial reference, i.e. only the name of places or objects such as Canada, Province of Quebec, Quebec City and St.John Street. This type of spatial dimension is the only one supported by conventional (non-spatial) OLAP tools. When used with SOLAP tools, the non-geometric spatial dimension is treated like the other descriptive dimensions and the geometric data allowing for the representation of the dimension members on maps is not used. In this case, the spatio-temporal analysis can be incomplete and certain spatial relations or correlations between the phenomena under study can be missed by the analyst. The two other types of spatial dimensions aim at minimizing this potential problem. To do so, the geometric spatial dimensions comprise, for all dimension members, at all levels of details, geometric shapes (ex. polygons to represent country boundaries) that are spatially referenced to allow their dimension members (ex. Canada) to be visualized and queried cartographically. The mixed spatial dimensions comprise geometric shapes for a subset of the members or the levels of details.

Fig. 1 presents an example of the three types of spatial dimensions [4].

Every level of a spatial dimension that has geometric shapes associated to its dimension members can support spatial drill operations (see Navigation operators: spatial drill) on cartographic features, thus increasing the number of degrees of freedom for interactive spatio-temporal exploration of data.

Spatial Measures:

Within a SOLAP tool, maps are used to display the members of the geometric or mixed spatial dimensions, using visual variables that relate to the values of the different measures contained in the datacube being analyzed. A SOLAP system also supports two types of spatial measures as well as spatial dimensions.

A first type of spatial measure is the set of all the geometries representing the spatial objects corresponding to a particular combination of dimension members (it is possible to have many geometric spatial dimensions in a datacube). It consists of a set of coordinates, which requires a geometric operation such as a spatial union, a spatial merge or a spatial intersection, to be computed. To implement this type of spatial measure, it may be necessary to use pointers (stored within the multidimensional data structure) to the geometric shapes stored in another structure or software.

A second type of spatial measure results from the computation of spatial metric or topological operators. Examples of this type of spatial measure could be “surface” and “distance” [3] as well as “number of neighbours”. Similarly, spatial dimensions can also be used to present the results of spatial analysis operations in a hierarchical manner (ex. adjacent–>adjacent by points–>adjacent by only one point) and can be used to find the facts that correspond to the selected spatial operator members of a spatial operators dimension [2].

The measure values are calculated by the OLAP or the SOLAP server. The server aggregates and physically stores the aggregations according to the possible combinations of dimension members. In the case of SOLAP servers, however, it is almost impossible to materialize all the possible geometric aggregations of spatial measures (or views) as this can result in an explosion of the necessary storage space. Algorithms have been defined in order to optimally select the spatial aggregations to be materialized.


References:

[1] Bédard, Y., Merrett, T., Han, J., 2001. Fundamentals of spatial data warehousing for geographic knowledge discovery. In: Miller, H., Han, J. (Eds.), Geographic Data Mining and Knowledge Discovery. Taylor and Francis, London, pp. 53–73.

[2] Marchand, P., Brisebois, A., Bédard, Y., Edwards, G., 2004. Implementation and evaluation of a hypercube-based method for spatio-temporal exploration and analysis. Journal of the International Society of Photogrammetry and Remote Sensing (ISPRS) 59 (1-2), pp. 6–20.

[3] Rivest, S., Bédard, Y., Marchand, P., 2001. Towards better support for spatial decision-making: defining the characteristics of Spatial On-Line Analytical Processing (SOLAP). Geomatica, the Journal of the Canadian Institute of Geomatics 55 (4), pp. 539–555.

[4] Rivest, S., Bédard, Y., Proulx, M.-J., Nadeau, M., 2003. SOLAP: a new type of user interface to support spatio-temporal multidimensional data exploration and analysis. Proceedings of the ISPRS Joint Workshop on Spatial, Temporal and Multi-Dimensional Data Modelling and Analysis, Quebec, Canada, October 2-3.

[5] Rivest, S., Y. Bédard, M.-J. Proulx, M. Nadeau, F. Hubert & J. Pastor, 2005. SOLAP: Merging Business Intelligence with Geospatial Technology for Interactive Spatio-Temporal Exploration and Analysis of Data, Journal of the International Society for Photogrammetry and Remote Sensing (ISPRS) "Advances in spatio-temporal analysis and representation" 60 (1), pp. 17-33.

 

Université Laval - Canada
Updated: November 2009