I'm designing my first data cube and wanted to entertain a discussion on which approach is best. Designing the schema first, or modelling the data once it's been imported into SAAS? The job I have, is that I'm in possession of a number of relational
databases, which the firm want to see "BI run against", using SSAS and Performance Point 2007. I've been given no granularity statement, so I'm left wondering, do I come up with a schema first, or do I try and do something with the data?
As I've never designed cubes before, please assume amateur level when posting any response. Any comments are welcome.
View Complete Post