Spatial Data Warehouses - Data Warehouse Systems: Design and Implementation - page 453

11.8.1 Requirements Specification and Conceptual
Design
Like for traditional data warehouses, spatial data warehouses can be designed
on the basis of the analysis requirements of the users, the data available
in the source systems, or both. This leads to three approaches for spatial
data warehouse design, referred to as the analysis-driven, source-driven, and
analysis/source-driven approaches.
An aspect that distinguishes spatial data warehouse design from conven-
tional data warehouse design is the need to define when we will consider
spatial support during the design process. If spatial support is considered in
the early steps of the requirements specification phase, we are in the case of
early inclusion of spatial support .Otherwise,weareinthecaseof late
inclusion of spatial support . As we will see, the choice between these
two options depends on the users' knowledge of spatial data features and
the presence of spatial data in the source systems. Further, this choice is
independent of which of the three approaches above is used. For example,
a particular data warehouse project might choose a source-driven approach
with early inclusion of spatial support in order to integrate existing spatial
applications into a decision-support infrastructure that is used by experts
cognizant of spatial databases.
Analysis-Driven Approach
In the analysis-driven approach to spatial data warehouse design, the
requirements specification phase is driven by the analysis needs of the users.
Figure 11.16 a shows the steps required in this approach. As can be seen in
that figure, in the first step of the requirements specification phase, users
at various management levels are identified to ensure that the requirements
will express the goals of the organization. These users will help the developer
team to understand the purpose of having a spatial data warehouse and to
determine the analysis needs, which are collected in the second step. The
information gathered and the corresponding metadata are documented in
the third step and serve as a basis for the next phase.
We next analyze how the two design choices for including spatial support
affect the requirements and conceptual modeling processes.
Early Inclusion of Spatial Support
In this case, we assume that the users are familiar with concepts related to
spatial data, including its manipulation and some kinds of spatial analysis.
Therefore, from the beginning of the requirements specification process, the