topic Re: Creating a performance index in SAS Programming

Creating a performance index

Deepakkala04 — Tue, 04 Feb 2014 14:00:43 GMT

I have a dataset with variables as "SalesID", "ActualSales" , "TargetSalesAmt" and "Region". I need to create a performance index based on total sales achieved against the target set, and identify the best performing region overall. How can this be done without using Proq SQL.

Example:-

SalsID	ActualSales	TargetSales	Region
XE908	19000	13800	East
XA908	20000	18000	East
XC102	25000	30000	West
XB102	15000	18000	East
XA908	20000	22000	West
XE908	10000	10000	North
XB102	30000	25000	South

Re: Creating a performance index

LinusH — Tue, 04 Feb 2014 16:14:33 GMT

Can you define in more detail what a performance index is? Perhaps by give some sample desired output?

And why not SQL?

Re: Creating a performance index

art297 — Tue, 04 Feb 2014 16:27:57 GMT

Depending upon your definition of performance index, something like the following might suffice:

data have;

infile cards dlm='09'x;

input SalsID $ ActualSales TargetSales Region $;

cards;

XE908 19000 13800 East

XA908 20000 18000 East

XC102 25000 30000 West

XB102 15000 18000 East

XA908 20000 22000 West

XE908 10000 10000 North

XB102 30000 25000 South

;

proc summary data=have nway;

var ActualSales TargetSales;

class region;

output out=want (drop=_:) sum=;

run;

data want;

set want;

growth=ActualSales/TargetSales-1;

run;

proc sort data=want;

by descending growth;

run;

Re: Creating a performance index

Deepakkala04 — Tue, 04 Feb 2014 16:30:44 GMT

Hi Linush.. Actually I want to represent the above data in a format which is some what like:-

1. Region with Highest Sale

2. Sales achieved (BY salesID) and also if they have been able to meet their targets

Y no SQL:- am looking for a very basic resolution for this problem. I want to test out if this is possible thru any basic SAS proc/data statements.(they might be a bit tedious)

Re: Creating a performance index

Peter_C — Tue, 04 Feb 2014 17:14:04 GMT

Merge against the output from

PROC MEANS data= dataset missing noprint nway ;

Class region ;

Var _numeric_ ;

Output out= index_perf( drop= _type_ )

Sum= /autoname ;

Run;

proc sort data= dataset ;

By region ;

Run ;

Data index_and_data ;

Merge dataset index_perf ;

By region ;

****** suitable performance calcs ;

run;