The business intelligence (BI) deployment provides the main elements DWH and data marts for collecting and storing data. But the OLAP is the key technology of BI and it is used for business analysis and decision support. According to [OMG02] the term OLAP is defined as follows:

“OnLine Analytical Processing (OLAP) is an analysis technique in which business data originating from multiple, diverse operational sources is consolidated and exposed in a multidimensional format that allows business analysts to explore it a retrieval-friendly environment” [OMG02, page 122].

With OLAP it is possible to aggregate multiple table queries faster than with normalised entities in an operational database [CBS02, page 943]. Based on these facts, a dimensional database is a better model for querying, but worse for operational use [WEnc04, OLAP].OLAP structured data enable both growth and trend analysis. The multidimensional format of OLAP data are very often represented by data cubes as displayed in figure 1. Such a data cube is divided into dimensions and measures appropriate to the DWH or data mart schemas in 2.2. In general, dimensions describes an analysis view of an application area that can be defined [BuG01, page 517]. In contrast, measures are the content (cell) of a data cube [BuG01, page 517] in dependency on a chosen analysis view (dimension). The dimensions are represented by every edge of the cube such as item, person and date in figure 1. A set of dimension values identify only one unique measure of a cube cell. For instance the dimension values person1001, item1277 and 05/12/2000 describe the measure actual revenue 342,82 in the corresponding cell as shown in figure 1.

Figure 1: Multidimensional data representation in a table a) with four arrays in comparison to a OLAP cube b) (adapted from [CBS02, page 945]).

Referring to the displayed OLAP cube in figure 1, it is possible to apply essential navigation operations for ad hoc analysis:

• Rotation/piovoting to new dimensional comparisons in the viewing area.
• Slice and dice to view different dimensions such as actual revenue per person by item.
• Rollup generates a result set showing aggregates for a hierarchy of values.
• Drill down/drill across enable to drill through from one level of detail to another.

Download article: OLAP_and_Reports.pdf