Factor Analysis in data mining

[JessB]

...Can it be done?

Basically, I've not been involved in remotely advanced mathematics since A-level. I'm now a marketing researcher and want to make my data mining efforts much more reliable for a new project. Ideally, I want to use factor analysis or PCT or something similar before getting started on discriminant analysis. The trouble is, any techniques I've found to reduce the number of factors requires that I do a survey (surely not the point of data mining...). I wondered if anyone had any ideas to start me off in my planning?

Any help much appreciated! :D