Aren’t end-users a pain? You provide them with a beautifully crafted cube in which all their business needs have been analyzed, dimensionalized, and sanitized and all they seem to do is complain that they can’t get the information they need. What are they on about? In this presentation I will discuss how report building imposes sometimes complex data modelling requirements that are likely to be lost in the traditional interview and analysis process and how capturing these requirements can lead to happier end-users. I will show some approaches to solving problems such as: 1. Handling of flow, balance and average data types 2. Providing commonly used time based aggregations (YTD, QTD, Rolling 12 months etc.) 3. Scaling of data to provide reports in Millions, Thousands and Units 4. Providing periodic and cumulative views of results 5. Allowing back-dating of reports so that the user can get correct results for prior reporting periods 6. Flexible handling of Scenarios and versions 7. And for those of you who get stuck with financial applications how to handle Credits, Debits and normal balance conversion The solution shows how addressing these reporting needs will lead to additional dimensions in the data model and the need to handle complex inter-dimensional relationships.
Presented by Andrew Wiles at SQLBits III