As a SAS Analytics Pro user, I mostly use SAS Bases's data steps, proc sql, proc freq, SAS/Stat's proc logistic. I use all these procedures in conjunction with a self-written library of SAS Macro code.
My industry is retail credit scorecards development.
I mostly use Proc sql, proc varclus, proc format, prof freq, proc summary, proc report, proc means, proc transpose, proc export and proc append. My industry is retail.
Primarily: Base, Macro, SQL, Stat.
Procedures: A lot of data steps, sql procedures, transpose, freq, univariate, summary, means etc...
Often things are sitting behind macro's though implying frequent use of call symput and symget, addressing dictionary tables, call execute, sysfunc, etc...
Industry: Pharma: Stat. Programming.
With ODS (layout) and Dashboards, Proc SGPLOT, Proc GKPI, Proc SGPANEL and Proc REPORT.
With Time series transformations into common time dimensions and derivatives, Proc Expand.
EM5.3 and now 6.2 is a must for PMML output and neural or logistic models (Try the rule induction for something unique - copyrighted by SAS themselves)
And of course the Powerdog Proc Means using Class variable.