Data
Warehouse
WHY DATA WAREHOUSE?
You wish to achieve the goals of your
company. You want to know your (potential) clients, their demands and needs. Furthermore
you want to know your present and future competitors, and what they are doing to meet your
clients needs.
Having established your goals, you want to
monitor them, in order to determine to what extent they are being achieved.
You do suspect that the management
information you need is available somewhere within your company, but you don’t know
where. When you ask your analysts, they come up with different results.
In Decision Making, managers have a
big business need for timely management information. It is vital to have a proper
structure in the whole compilation of data within your company. In many organizations
however, the current structure blocks effective and efficient use of existing data. Most
problems are caused by the fact that the required information is stored in different
systems, using different definitions or different formats. This frustrates a consistent
overall view on how your company is doing.
The creation of an integrated source of data
may enhance the performance of your decision support system, being specifically designed
to provide information in support of the decision making process, business wide analysis
and performance monitoring. This is what we call a data warehouse.
Characteristics of a Data Warehouse:
- Subject Oriented: data is organized around the subjects of
interest to managers, such as clients, markets, products, suppliers, etc.
- Integrated: data from different sources is harmonized and
uniquely coded;
- Time Dependent: all data is stored together with a date/time
stamp;
- Consistent: data is uniquely defined, its description
(location, meaning, ownership) is stored in a metabase;
- Nonvolatile and Historical Integrity: data is periodically
refreshed, re-aggregated as required, but in principle not changed.
By means of a data warehouse, it helps making available the
information you need.
The
Benefits of a Data Warehouse:
- Data derived from different sources can be retrieved in an
integrated fashion;
- Due to the subject oriented design, data can be approached
from (a combination of) different angles or dimensions, enabling multidimensional
analysis.
- Reports are reliable and verifiable: you are able to search
through the metabase to determine the source of the data and the rules which were used to
produce it;
- Analyses produced in different parts of the organization are
consistent and are reproducible;
- The number of interfaces to be maintained between transaction
processing systems and decision support systems is minimal;
- Due to the fact that historical data is maintained, it
becomes possible to carry out analyses over periods of time and to do datamining;
- Because transaction processing systems are separate from the
data warehouse, it becomes possible to perform complex analyses without affecting the
performance of these systems
Data Warehouse Architecture
Links:
Unisys
Data Warehouse description
Business Intelligence & Data Warehous http://www.dmreview.com
|