Business Intelligence / Data Warehouse
|
|
|
|
Business Intelligence (BI) and Data Warehousing (DWH) are not projects that are being defined, implemented and completed. It is an ongoing process and must not only be deeply rooted in the corporate culture but also in harmony with other corporate processes. Those that take the above principle to heart will be successful in implementing BI/DWH.
Originally, BI was a term used to describe the content and commercial aspects of a process while the term DWH was used to describe technological and database related aspects of the same process. Recently, two technological categories have developed as manufacturers have divided their products into these categories.
|
BI/DWH Pyramid |
|
„BI/DWH is a process deeply rooted in the corporate culture
and in harmony with the remaining corporate processes.“
Alexander Jochum, DATA MART
|
|
|
One determines the other (BI ↔ DWH)
|
It is quite adequate to compare the entire process with an iceberg. Only a fraction of the entire structure (10 - 25% of expenses) is visible in the end: BI. The largest part (DWH) and the largest expense remain under water and invisible to the user.
A Data Warehouse process ensures access to and a periodic and automatic transfer of data from a number of operative or even external sources. The first ETL process (Extraction, Transformation, Loading) must ensure data quality. It will not only perform plausibility checks on data arriving from very diverse sources but also provides data with structural information allowing a combination of data from different sources. This process transforms data into new information for the first time.
In essence, the next step aims at transforming structured data into new information, ranging from structured performance figures (KPI) to automated, rule-based data mining.
An important prerequisite for the acceptance and thus the success of a BI/DWH system is missing if the design of this process lacks in method, expertise, content or technical implementation: correct data and information.
What’s the use of perfect information if it is not user-friendly and not prepared with sufficient intelligence for the user? After all, structured information is supposed to reveal complex relationships and expose weak spots and unused potentials. These requirements are at the core of business Intelligence and a “must have” for BI tool selection. |
BI/DWH Iceberg |
|
„It may be quite sensible to provide different groups of users with different but applicable BI tools for
analysis or reporting - last but not least from a cost/benefit viewpoint.“
Christiane Breuer, DATA MART
|
|
True - much of a BI interface is a matter of taste. But it must live up to the requirement profile.
- For instance, a multi-dimensional OLAP database may but not necessarily has to be used;
what are the pro and cons in regards to requirements?
- Is a high-density cockpit necessary and sensible for all user groups?
- When do we put our money on tables and when do we use graphics or a mix of both?
- Who may see what?
These and many more questions have to be asked and answered interdisciplinary across all hierarchy levels and departments. Only then did we achieve the second ingredient of success of a BI/DWH system, quite in addition to the correctness of data and information: the commercial contents arrive at the user according to requirements. | |
|
|
|
|
|