Most delegates will be familiar with the many proprietary tabulation systems available. When we set out to build Vector, a web based cross tabulation system, we broke the problem down into the two classic phases, aggregation and manipulation. However, we decided that, rather than developing yet another proprietary set of code for table manipulation, we would utilize a mainstream standard, MDX (Multi-Dimensional eXpressions), which is utilized in OLAP within Microsoft Analysis Services (part of SQL Server), as well as by other major vendors. While OLAP may not seem immediately obvious in application to survey analysis, it permits the manipulation of aggregated data, which is a necessary phase in all but the most simple of cross-tabulation requirements.
This paper will examine the functionality of MDX and the applicability of the language to survey analysis, and decisions that were made to cope with hierarchical data and user-derived variables. We will also present our thoughts on XMLA, a new open standard for OLAP systems.
| Back to: Top | Programme | Page last updated on 31 August, 2003 |