topic Re: Creating Multiple Graphs Against Common Y Axis in ODS and Base Reporting

Creating Multiple Graphs Against Common Y Axis

deleted_user — Mon, 07 Jan 2008 20:55:31 GMT

Hi all,

I currently have 10 explanatory variables, and I would like to compare each one against 1 common dependent variable, separately (which would require 10 different scatterplots).

My current process is pretty tedious, it involves the finding the average of the dependent variable for each occurence of each explanatory variable (sample code below)

PROC SQL;
CREATE VIEW LINE_SUMMARY1 AS
SELECT b1_last12,
AVG(BAD) AS BAD_AVG
FROM WORK.OUTCOME2
GROUP BY b1_last12
ORDER BY b1_last12;
QUIT;

Then plotting out the explanatory variable against the average outcome of the dependent variable

PROC GPLOT DATA = LINE_SUMMARY1
;
PLOT BAD_AVG * (B1_last12) /

RUN;

As it stands, I basically have to run these portions 10 times (for each explanatory var.). Is there some way to simplify this process? Thanks!

Re: Creating Multiple Graphs Against Common Y Axis

Olivier — Mon, 07 Jan 2008 21:53:13 GMT

Hi.
You can compute all means at the same time, then use only one Gplot procedure to produce all graphs.[pre]
ODS EXCLUDE ALL ; /* prevents any printing of statistics */
ODS OUTPUT summary=work.means ; /* saves means in a dataset */
PROC MEANS DATA = work.outcome2 MEAN ;
VAR bad ;
CLASS b1_last12 /* and other explanatory variables */ ;
WAYS 1 ; /* compute means by each explanatory variable */
RUN ;
ODS SELECT ALL ; /* results now get printed as usual */
PROC GPLOT DATA = work.means ;
PLOT bad_mean * (b1_last12 /* and other explanatory variables */ ) ;
RUN ; QUIT ;
[/pre]
Regards.
Olivier

Re: Creating Multiple Graphs Against Common Y Axis

deleted_user — Tue, 08 Jan 2008 13:05:05 GMT

good point, thanks alot!

:-)