Information Technology Reference
In-Depth Information
Table 1.
Identified conceptual design approaches for DW modelling
Approaches
Description
A Goal-driven approach [2, 6, 8, 9, 10] places emphasis on the need to align data
warehouse with corporate strategy and business objectives (discovering goals of deci-
sion makers). The requirement analysis stage aims at obtaining informational require-
ments of decision makers. These requirements are related to interesting measures of
business processes and the context for analysing these measures. These set of visions
must converge in order to obtain a set of qualified KPI.
This approach works only well when business processes are designed throughout the
company and are combined with business goals.
Goal-driven
This modelling approach (also known as Demand-driven) [2, 4, 9, 11] is carried out
based on business-users requirements to generate the multidimensional schema. A
User-driven approach is mainly focused on requirement analysis process and assumes
that the requirements are exhaustive and deal with approaches facilitating user partici-
pations. As such, they do not consider the data sources to contain some interesting
elements for analysis. This means that information requirements are mapped with the
available data sources only a posteriori, sometimes this mapping fails and may cause
users' disappointment.
User-driven
In this modelling approach (also known as Supply-driven) [2, 4, 9, 12] the multidi-
mensional model for the data warehouse is derived by thoroughly analysing the data
source. As a consequence, the supply-driven rely on discovering as much multidimen-
sional knowledge as possible from the data sources. This approach has a disadvantage
to generate too many results, which misleads the user and may never be used. Re-
quirements are used at the end of design to filter the results.
A detailed comparison of data-, goal-, and user-driven can be found in [6]. This article
concludes that the three methods are complementary and should be used in parallel to
achieve optimal design.
Data-driven
The principles of this approach [8, 9] are based on the relationships between business-
processes and Entity-Relationship-Models. The aim is to define (or redefine) the or-
ganizational processes (as-is) which will supply the DW data and to ensure that they
are aligned with the business goals (to-be). The added value of this approach will be
the integration of the previous methods (user-driven and data-driven) with organiza-
tional processes that will treat these sets of information's to be used. Finally, the goal-
driven will verify if the process-driven achieve the business goals.
Process-driven
The data warehouse design process [8] is carried out by extracting data warehouse
ontology from global ontology and annotating the concepts with multidimensional role
by analysis of defined goals of the requirement model. For each business requirement
annotated subset ontology is derived from source ontology and validated for multidi-
mensionality. Since ontology can capture the concepts of a domain in a formal and
meaningful way, each data source of the data warehouse can be represented using
ontology.
Ontology-
driven
This approach [8, 9] is complementary to the other approaches and is focused in iden-
tifying the existing technological capability in the organization to implement the DW.
This result in a diagnosis of the current state of the technology in the organization and
the level to which organizational processes and the type and size of DW can be im-
plemented and grow with the technological capability identified. This diagnosis aims
at finding solutions to mitigate the impact of technological constrains in designing the
DW conceptual model.
Technology-
driven
Search WWH ::
Custom Search