| Data warehousing and OLAP (online analytical processing) technologies have gained a widespread acceptance since the 90’s as a support for decision-making. A data warehouse is a collection of subject-oriented, integrated, consolidated, time-varying, and non-volatile data (Kimball, 1996; Inmon, 1996). It is manipulated through OLAP tools, which offer visualization and navigation mechanisms of multidimensional data views commonly called data cubes.
A data cube is a multidimensional representation used to view data in a warehouse (Chaudhuri & Dayal, 1997). The data cube contains facts or cells that have measures, which are values based on a set of dimensions where each dimension usually consists of a set of categorical descriptors called attributes or members. Consider for example a sales application where the dimensions of interest may include, costumer, product, location, and time. If the measure of interest in this application is the sales amount, then an OLAP fact represents the sales measure corresponding to a single member in the considered dimensions. A dimension may be organized into a hierarchy. For instance, the location dimension may form the hierarchy city state region. Such dimension hierarchies allow different levels of granularity in the data warehouse. For example, a region corresponds to a high level of granularity whereas a city corresponds to a lower level. Classical aggregation in OLAP considers the process of summarizing data values by moving from a hierarchical level of a dimension to a higher one. Typically, additive data are suitable for simple computation according to aggregation functions (SUM, AVERAGE, MAX, MIN, and COUNT). For example, according to such a computation, a user may observe the sum of sales of products according to year and region. |
