@Cruise wrote:
@novinosrin
the data HAVE is posted in my original post. It has in total around 900 ZIP codes. that's it. not a large data.
Well, 1449 to be more precise (see my PROC FREQ output).
(I assume that the table of ZIP code categories with columns PUP, VAT, CAT and OTHER [and a ridiculous number of empty cells due to unequal sizes of the four categories] is in fact a spreadsheet, not a SAS dataset.)