Evaluation criteria

Many commercial products can be used to deploy a solution that combines multidimensional analysis and cartographic visualization. Various levels of functionality exist, but only a few products offer the architecture and functions required to support full SOLAP capabilities.

In the Summer of 2007, the Industrial Research Chair in Geospatial Databases for Decision-Support team from Université Laval (Quebec) established a number of evaluation criteria [2] to distinguish commercial products that support Spatial OLAP capabilities. These evaluation criteria can be used to identify the products that will best suit your needs.

In order to do so, you have to identify your preferences for each of the criteria described in the evaluation grid. In certain cases, a given preference can have a great influence on the final choice, for example, if the architecture has to support a particular spatial data format or if the solution does not require an OLAP Server. You can then compare the preferences with the studied solutions. The evaluation criteria are grouped in two categories: criteria related to the supported architecture and criteria related to the supported features.


1- Evaluation criteria related to the supported architecture

Supported cartographic component

In fact, many cartographic tools can be used to build the cartographic component of a SOLAP solution, such as GIS technologies (e.g. ESRI ArcGIS), WEB-GIS (e.g. MapInfo MapXtreme, ESRI ArcIMS), cartographic viewers (e.g. ESRI ArcExplorer), vector viewers (e.g. SVG viewer) or visualization tools (e.g. AVS). Some of these cartographic components support the popular spatial data formats that are commonly used in the geomatics industry.

If the tool supports recognized and standardized spatial data formats (e.g. ESRI Shapefile, MapInfo MID/MIF), the management of the spatial data and their updates in the application are simplified.

The studied criteria are: cartographic component type, cartographic component name and supported spatial data formats.

Supported multidimensional component

The multidimensional component can be an OLAP server (e.g. Microsoft Analysis Services, SAS OLAP, SAP BW), an OLAP client (e.g. MicroStrategy, Business Objects XI), a disconnected access to a multidimensional database (e.g. Microsoft Analysis Services cube files (.cub)) or an access to a star schema deployed using a relational database.

When an OLAP server or an OLAP client is required in the solution, the datacube management is facilitated with a dedicated data structure, specialized indexes, powerful aggregation algorithms, etc. On the other hand, the option to use an OLAP server/client makes the solution more expensive to implement.

The studied criteria are: name of the supported OLAP server/client (if needed) and name of the supported multidimensional access.

Coupling approaches

Implementing a SOLAP system requires major developments in order to build a custom front-end and to use OLAP and cartographic components as back-end services. According to LGS Group [1], various basic approaches to GIS and OLAP tools integration can be considered: GIS-dominant, OLAP-dominant and total integration.

Integrated SOLAP technology is still uncommon in the commercial market and most often, integration is still partial between the two components.

The GIS-dominant approach offers GIS capabilities inside a GIS graphical user interface. Usually, only simplified access to OLAP data sources is offered and practically no Spatial OLAP capabilities (e.g. drill on maps) are supported.

The OLAP-dominant approach offers full OLAP capabilities inside an OLAP graphical user interface, but limited GIS capabilities (some tools offer spatial selection of objects). Generally, these solutions are bidirectional between the OLAP and the GIS, which implies that the selection on one hand (e.g. on the map) affects the results on the other hand (e.g. in the OLAP table).

In an integrated solution, both the OLAP and the GIS features are offered. The main graphical user interface is custom and built over the underlying technologies (i.e. OLAP and GIS) and the SOLAP application is developed to benefit from the integration of OLAP and GIS functions. These tools integrate the management of the multidimensional and spatial data and they facilitate the deployment and the update of spatial datacubes.

The studied criterion is: Type of coupling approach (OLAP-dominant/GIS-dominant/Integrated solution).

Table 1 presents the evaluation grid comprising the criteria related to the supported architecture.

 


2- Evaluation criteria related to the supported features

The aspects related to the features supported by the solution compose the second set of evaluation criteria.

Supported spatial datacube structure

In order to support drill operations on the spatial levels of a spatial dimension, a link must be defined between the multidimensional data and the spatial data. By linking the spatial members of a dimension to the corresponding geometries stored in a spatial database, these members can be viewed on a map. Various types of spatial dimensions can be supported in a SOLAP application. These concepts are explained in the Concepts page of this site.

It is necessary to configure a SOLAP application to support these concepts. This configuration can be assisted (using wizards) or left to the administrator (ex. using XML). The level of user assitance in these operations is an evaluation criteria.

For this point, the studied criteria are: The level of spatial concepts supported (spatial member/spatial level) and the level of user assitance (Y/N).

Supported spatial drill operators

The spatial operators are spatial drill on members, spatial drill on selections or spatial drill on levels. They are executed directly by clicking on the elements (dimension members) shown on the maps or on a selection of many elements (at the same or at different levels of details of a dimension). The spatial operators that are supported by a SOLAP solution are illustrated in section "Exploration operations available for the map displays" of the Exploration page of this site. Usually, commercial solutions do not support all of these levels of interaction.

For this point, the studied criterion is: The level of supported spatial drill operators (by member/by selection/by level).

Advanced maps supported

Various types of maps can be supported by a SOLAP solution. Simple maps are common. Multimaps, complex thematic maps and maps with superimposed diagrams are complex maps that can be used to represent many dimensions or many measures on a map. The types of maps that can be supported by a SOLAP solution are illustrated in section "Many view types" of the Visualization page of this site.

For this point, the studied criterion is: Advanced maps supported (Y/N).

Intelligent automatic map creation

The map creation process may be done manually by the user. But the creation of a simple thematic map can be a hard task when the map has to conform to graphical semiology rules. SOLAP tools can use built-in rules to generate coherent maps based on the user's selections. For example, on a map generated after a drill down operation, the SOLAP tool can automatically apply the appropriate classifications and colors by following predefined display rules. The use of such automation rules helps enhance the knowledge discovery process and helps the user to maintain his(her) train of thoughts.

For this point, the studied criterion is: Automatic map creation (Y/N).

Pivoting dimensions on maps

The pivot is a useful operation on multidimensional data. Applied on a map, the pivot allows to produce a different type of map. However, to be successfully applied, precise rules are necessary to produce the pivoted map that corresponds to the active dimension selections. For example, a multimap can be pivoted as a map with superimposed diagrams. For more information on pivot operations on maps, see the Exploration page.

For this point, the studied criterion is: Pivoting dimensions on maps (Y/N).

Table 2 presents the evaluation grid comprising the criteria related to the supported features.


 

References:

[1] LGS Group, 2000. Analysis of health surveillance business intelligence tools & applications. Final report for Health Canada.

[2] Proulx, M.-J. & Y. Bédard, 2007, Evaluation of commercial products that offer combined capabilities of OLAP analysis and cartography. (French only - Évaluation des produits commerciaux offrant des capacités combinées d'analyse OLAP et de cartographie), Research report, Industrial Research Chair in Geospatial Databases for Decision Support, Department of Geomatics Sciences, September 25th, 2007, Université Laval. 55 pp.

 

Université Laval - Canada
Updated: November 2009