Database Reference
In-Depth Information
shown in Fig. 11.16 b, prior to the development of the final schema and the
corresponding mappings, there is an additional step to add spatial support.
In this step, the designers present the conceptual schema to the users and
ask them for indications about the spatial support required.
If the MultiDim model is used as a conceptual model for designing a
spatial data warehouse, in the first step the designers may consider each
level and decide whether that level, some of its attributes, or both should
be represented spatially. Then, if a hierarchy includes two related spatial
levels, a topological constraint between them may be specified. If a fact
relates two or more spatial dimensions, the designers can help the users to
determine whether a topological constraint exists between these dimensions.
Finally, the inclusion of spatial measures may be considered. Note that the
elements of the multidimensional schema could be analyzed in a different
order, depending on the designers' skills and their knowledge about spatial
data warehouses and the particularities of the conceptual model used. Similar
to the previous case, the step of checking data availability may require access
to external sources since spatial data may not be present in the underlying
source systems. The final schema should include the modified mappings.
As an example, consider the schema in Fig. 4.2 , developed for a traditional
data warehouse following the method described in Chap. 10 . When this
schema was shown to the users, they required the possibility to visualize
on maps the geographic hierarchies for Customer , Supplier ,and Employee
dimensions. Figure 11.4 shows the addition of geographic properties to the
hierarchies in the initial conceptual schema. We use the spatial extension of
the MultiDim model described in Sect. 11.2 . The spatial elements are then
checked against the data available in the source systems. Since the operational
data do not include spatial components, external sources were used to obtain
the corresponding information for these spatial hierarchies.
Source-Driven Approach
As explained in Chap. 10 , this approach relies on the data in the source
systems. Like in the analysis-driven approach, spatial support may be
included either early or late in the design process. Since the operational
databases are the driving force in this approach, the choice between early
or late inclusion of spatial requirements depends on whether these databases
are spatial or not.
Early Inclusion of Spatial Support
If the source systems include spatial data, steps similar to those for traditional
data warehouse design can be applied. They are indicated in Fig. 11.17 a
and the lower path of Fig. 11.17 b. Requirements specification starts with
Search WWH ::




Custom Search