Multidimensional OLAP Analysis
Multidimensional analysis is a concept which lets users evaluate multiple aspects of the business against company's key performance measures.
It is an easy way to understand the business data, decrease the time needed to get familiar with the data and increase quality of the decisions made by people who make business decisions.
Browsing through the multidimensional data in an unlimited combination of views may help uncover successes and failures, find opportunities and get to know trends.
Multidimensional OLAP cube is a fundamental data structure in every OLAP system working in a DataWarehouse environment.
When the data is organized into multidimensional structure, it allows users to explore and answer questions related to the organization and helps realize what is driving the business and measure performance accross various levels.
A cube consists of Measures, Dimensions and Levels and is optimized for secure and fast retrieval of multidimensional data.
Dimensions represent descriptive data about the major aspects of the system (who? what? where? when?). Dimensions may be referred to as a series of viewpoints and perspectives related to the organization.
Dimensions are groupped into levels that represent the hierarchy set up by an organization and enable users to increase or descrease the amount of details related to a dimension.
Measures are related to the numeric indicators (quantitative data) that are collected and stored in a company and that measure the performance of the business (they answer the questions like how much? how many?).
A graphical representation of multidimensional OLAP cube can be seen when exploring Microsoft Excel Pivot Tables (crosstabs) or Cognos PowerPlay cubes (PowerCubes).
Usually a DataWarehouse system have many different dimensions specific to the given organization, however the dimensions mentioned below are met most often:
A hierarchy of each of the dimensions can be established by setting up Levels.
For example, level hierarchy can be set up in the following way:
Categories are data items that populate levels in a dimension. For instance, let's take into consideration the Location dimension whew the levels set up in a Data Warehouse system are: Area, Region, Country, City.
Let's consider the city of Quebec and its drill-down path which might be set up in the following way:
Measures can be derived directly from the data source (Net sales for instance) as well as calculated (Profit Margin = Net sales-Costs-Rebates&Bonus).
A list of measures which typically exist in a DataWarehouse system and are included in most of the OLAP cubes and reports: