Business Intelligence: Multidimensional Data Analysis

The relational database model is probably the most frequently used database model today. It has its strengths, but it doesn’t perform very well with complex queries and analysis of very large sets of data. As computers have grown more potent, resulting in the possibility to store very large data volumes, the need for efficient analysis and processing of such data sets has emerged. The concept of Online Analytical Processing (OLAP) was developed to meet this need. The main OLAP component is the data cube, which is a multidimensional database model that with various techniques has accomplished an incredible speed-up of analysing and processing large data sets. A concept that is advancing in modern computing industry is Business Intelligence (BI), which is fully dependent upon OLAP cubes. The term refers to a set of tools used for multidimensional data analysis, with the main purpose to facilitate decision making. This thesis looks into the concept of BI, focusing on the OLAP technology and date cubes. Two different approaches to cubes are examined and compared; Multidimensional Online Analytical Processing (MOLAP) and Relational Online Analytical Processing (ROLAP). As a practical part of the thesis, a BI project was implemented for the consulting company Sogeti Sverige AB. The aim of the project was to implement a prototype for easy access to, and visualisation of their internal economical data. There was no easy way for the consultants to view their reported data, such as how many hours they have been working every week, so the prototype was intended to propose a possible method. Finally, a performance study was conducted, including a small scale experiment comparing the performance of ROLAP, MOLAP and querying against the data warehouse. The results of the experiment indicates that ROLAP is generally the better choice for data cubing.