Proc SGPanel not displaying expected results (using sashelp.cars)

[Buzzy_Bee]

I was trying to create histograms for the sashelp.cars data set, and I want price (MSRP) binned into 5 groups and I also want to display Origin and Type on the plot. However, my resulting display just isn't right and I'm not sure what I'm doing wrong. A quick proc freq shows that Sedans are by far the most common Type of car in the three car manufacturing Origins, but the stacked plot shows USA mostly has Sports cars. Sports is the least common in all regions so it seems my charts are somehow backwards.

I don't have to use SGPanel if there is something better that you recommend, but I do want to be able to bin my continuous variable (price in this case). Thanks for your help 🙂

proc freq data=sashelp.cars; 
tables type*origin;
run;
proc sgpanel data=sashelp.cars;
	where type in('SUV','Sedan','Sports');
	panelby origin;
	histogram msrp / group=type nbins=5;
run;

[ballardw · Accepted solution]

See if this makes more sense.

proc sgpanel data=sashelp.cars;
	where type in('SUV','Sedan','Sports');
	panelby origin;
	histogram msrp / group=type scale=count nbins=5;
run;

The default for the scale is percent and with grouped values I think the denominator is not all cars but the type of car. So roughly 37 percent of Sports cars are in the bin with the price of $50,000

[Buzzy_Bee]

Perfect - that's exactly what I was looking for. The example I was following on the documention.sas.com website showed their option as "stat=percent" so I had tried changing it to stat=freq, but kept getting error messages and I couldn't find better examples on the documention.sas.com website. Your solution using scale=count is the correct option to use to get it display how I wanted it. Thank you.