An Idea Exchange for SAS software and services

Comments
by SAS Employee RMP
on ‎12-07-2013 04:13 AM

not quite sure I understand the question. Surely data modeling is an activity which would happen prior to using any of the above SAS tools/methods?

by Respected Advisor
on ‎12-09-2013 05:56 AM

In my experience you're using EG to prep the data for modelling during prototyping. To operationalize things you're using DIS to create the data mart and you run the operationalized models against the data created using DIS. This process is very likely scheduled using LSF or whatever is the customer's enterprise scheduler.

Ongoing new modelling will still use EG for data prep, eg. for enriching the existing operationalized data mart - and if successful new models require additional columns then a change is required using a full SDLC process.

by SAS Employee anna_holland
on ‎12-10-2013 11:43 AM

Hi RMP ,

You are correct. Data modeling is an activity that does happen before any other SAS tools/ methods are applied. Thank you, Patrick, for answering that question in great detail. I have re-worded the question to emphasize tools that manage and process data.

Thanks,

-Anna-Marie

by Respected Advisor
on ‎12-10-2013 03:15 PM

Hi Ana-Marie

You're question suggests now "multi choice" but you only allow a single choice.

Also: It can be very useful to use DataFlux for data profiling already during the data modelling process.

Thanks

Patrick

by Occasional Contributor BrianDoesSas
on ‎01-15-2014 07:42 PM

What Patrick said.

For "routine processing" you should (imho) be using DI if you have it. And by DI, I mean, actually putting things in DI blobs with metadata around it.

EG for exploration.

Idea Statuses
Top Liked Authors