Hi

I'm working on a project to gather feedback from our customers. The survey
contained 50 questions. Now we have the results we need a way to analyze
the answers. The team analyzing the data love analysis services and the
cubes we have currently created for them. So they would like to see a cube
with all the 50 answers (some would be dimensions, some measures, some
both).

I'm guessing a 50 dimension cube would kill analysis services? Does having
multiple smaller cubes linked with a virtual cube work better?

Any suggestions?

"Paul" <pe (AT) nospam (DOT) com> wrote

I've looked at this before and I know some people do use Analysis Services
successfully for questionnaire analysis. The killer is if you have a lot of
multi-choice questions where multiple selections can be made -- each choice
ends up as a dimension and you can get into hundreds of dimensions in no
time. But 50 dimensions isn't too bad in Analysis Services, particularly as
the volumes are likely to be very small by data warehouse standards and the
calculations quite simple.

[As a sideline, we are currently analysing the fascinating data for The OLAP
Survey 3 and, sadly, SPSS has to be used to do the crosstabs as there would
be far too many dimensions for any OLAP tool, including Analysis Services.]

Nigel Pendse
OLAP Solutions
http://www.olapreport.com