|
People with OR backgrounds
are likely to have a number of skills that are pertinent to
data warehousing, including:
- extracting and interpreting data;
- understanding quantitative analysis;
- working in and managing multi-disciplinary project teams;
- taking an enterprise viewpoint;
- managing change;
- designing and implementing effective decision support
systems;
- modelling business processes.
More specifically, OR people are well-placed to contribute
to data warehousing initiatives by helping the project team
to:
- take a rational approach (e.g. cost/benefit analysis)
to project justification;
- identify and prioritise the deliverables;
- probe and document the meaning and quality of available
data;
- decide when data are clean enough for their intended
purpose;
- choose useful and appropriate performance measures;
- make effective use of prototyping;
and by:
- facilitating cross-functional collaboration and compromise;
- bridging the gaps between business and technical staff;
- championing the use of analytic tools and thinking to
improve decision making;
- proposing and evaluating process changes that make better
use of information;
- developing models to populate parts of the data warehouse
with calculated data e.g. forecasts or allocated costs;
- pioneering the integration of new data sources with existing
databases;
- designing decision support systems and analytic
applications based on data in the warehouse e.g. demand
forecasting.
Clearly OR has much to offer in making a data
warehouse successful.
 How
can a data warehouse help OR? |
Once a data warehouse is available, OR stands to gain in
at least three ways:
1) Availability of useful data
Detailed data, in greater volumes and from many different
sources, should be much more readily available. This should:
- allow greater accuracy;
- support the use of more sophisticated techniques, including
statistical analysis and data mining;
- increase the range of problems that can be addressed.
2) Data preparation effort
Substantially less time is likely to be needed to collect,
understand, clean and check data used for modelling and
analysis. Furthermore, data from more different sources
are likely to have been linked on a consistent basis already,
leading to even greater time savings. Also, if large volumes
of operational data have been captured automatically, sampling
may no longer be needed.
3) Organisational climate
The construction of a data warehouse is likely to reflect
increased management interest in quantitative analysis,
and may lead to wider use of basic analysis techniques by
non-OR staff and greater acceptance of more "rational"
approaches to decision making.
The proliferation of data warehouses is creating opportunities
in many different organisations to apply more sophisticated
techniques such as statistical analysis and data
mining, and to develop analytic applications in new areas
such as supply chain integration, customer
relationship management and e-business.
For more information on the relevance of data warehousing
to these areas, see other
disciplines.
|
