Bart Simpson and Regression, Part II

Here at GVSUG, we continue to promote the most cutting edge and most useful SAS tools for our members. So, with great pride, today we present part 2 of our series The Use of Bart Simpson Images in Linear Regression.

Jeff Cartier of SAS Institute today has provided us with clear proof that our earlier claims about SAS/GRAPH were incorrect. We claimed that while Excel could include a Bart Simpson image as background for a linear regression plot, SAS/GRAPH could not. Cartier’s graphics and SAS code show that SAS/GRAPH is indeed capable of this state-of-the-art statistical graphic.

Here is Jeff’s code:

goptions reset=all dev=gif ftext="Akbar"
ftitle="Akbar/bo" htext=14pt htitle=18pt
vsize=5in hsize=3in;
title "Bart likes Linear Regression";
footnote j=r "Produced with SAS/Graph(r)";
axis1 minor=none;
axis2 label=(angle=90) minor=none;
symbol1 value=diamond i=rlclm95 width=2 height=2;
proc gplot data=sashelp.class;
plot Weight*Height / haxis=axis1 vaxis=axis2 autovref
iframe="c:\images\bart.png" imagestyle=tile regeqn;
run;
quit;

The Akbar font can be downloaded here. The background image of Bart Simpson is here.

But wait just a cotton-picking minute Jeff, where’s the R-squared value? Excel could do that!

Best Statistical Graph Ever!

Or should it say “Worst Statistical Graph Ever”?

This graph was taken from an actual e-mail I received. Prior to seeing this graph, I would have said that Excel is never better than SAS for statistical analysis, however, this graph clearly illustrates a capability of Excel which is superior to SAS.

GVSUG Meeting March 28

The next GVSUG Meeting will be March 28 at 3:30 pm at Brighton Memorial Library. We will have four talks, a networking session, an “Ask The Experts” session and refreshments. The entire agenda is here. The four talks are:

• Keynote talk: “Contrasting programming techniques for summarizing voluminous SAS output using the SAS Output Delivery System (ODS) (PROC FREQ as an example)” by Stuart Long, Westat
• “Where to Get SAS® Help When the Local Goo-Roo Is Out” by Earl Westerlund, University of Rochester
• “Macros that do Windows, Data and other tricks” by Michael E. Tomb, Informatica ECS, Inc.
• “Univariate Outlier Detection Using a Macro based on the IML Routines LTS and LMS” by John Hixon, Eastman Kodak Company

Registration will be from 3pm to 3:30pm. If you paid your dues at our October meeting, you don’t need to bring additional dues, because you paid for a whole year. Otherwise, we charge $15 for regular members and $10 for students, retirees and unemployed. Checks or cash only. Everyone (whether you have to pay dues or not) should pre-register by filling out the registration form.

Business Intelligence SAS Users Group

BISUG is a special-interest group serving the North American and European SAS communities. The main goal of BISUG is to promote the sharing of information related to SAS Business Intelligence products, including Enterprise Guide, Add-In for Microsoft Office and Web Report Studio, in the SAS user community.