Hello, I am programming boxplots using proc boxplot with a grouping variable (trt)
The code goes like this:
proc boxplot data = test3 ;
symbol value=dot height=1 c=blue;
plot var1* time (trt) / cframe = vligb
cboxes = dagr
cboxfill =ywh ;
My question is whether there is a way to fill the boxes with aternative colors - since I am comparing 2 treatment groups? The x-axis is arranged such that for for time=1 I have 2 boxes above - one for trt A and one for trt B. After that there is time=2 with 2 boxes above - A and B. I want to differentiate between A and B.
Thanks in advance.
If none of the built-in boxplots (proc boxplot, gtl, or gplot with interpol=box) support this, you might have to use some custom programming to create it.
Here's an example I wrote, that draws basic boxplot geometry with annotate - it would require quite a bit of further customizing to get the exact boxplot you're wanting. You could programmatically control the color of each box (based on values in the data), etc. I guess the 'tough' part would be visually/spatially "grouping" the bars together, since the annotate coordinates are based on gplot, and gplot doesn't have support for official groups.
Here's the code:
I describe how the code works, in detail, in my book "SAS/Graph: Beyond the Basics" - it's Example 10, on pp. 65-70.
Your code looks quite complicated. It will take me some time. I was hoping for a more "built in" solution. Is there a way to use annotate (with which I'm really not very familiar ) to color every other box? And can I use annotate with proc boxplot?
The answer depends on which release you are using.
At SAS 9.2 - A simple (non-grouped) Horizontal box plot can be made as below:
proc sgplot data=sashelp.cars;
hbox horsepower / category=type;
At SAS 9.3 - A grouped Horizontal box plot can be made as below:
proc sgplot data=sashelp.cars;
hbox horsepower / category=type group=origin groupdisplay=cluster;
But, you may not have SAS 9.3, in which case you can use SAS 9.2 GTL by overlaying two box plots, and use DiscreteOffsets to position them side by side. See this article on Distribution of Max LFT.
hi ... here's a 'non-SG solution' ...
I made up some data similar to yours (sort of, maybe you can tweak your data a little to match what I used)
I used the CBOXFILL=(variable) option to change colors according to treatment ...
array color(2) $4 _temporary_ ('red' 'blue');
do time=1 to 5;
do trt=1 to 2;
* vary color by treatment;
* 100 observations at each time/treatment combo;
do k=1 to 100;
var1 = ceil(trt*500*ranuni(0));
proc sort data=x;
1='TIME 1' 2='TIME 2' 3='TIME 3' 4='TIME 4' 5='TIME 5';
goptions reset=all gunit=pct ftext='calibri' htext=2;
axis1 value=none label=none major=none ;
axis2 label=('VARIABLE 1');
* white space around plot;
title2 a=90 ls=2;
title3 a=-90 ls=2;
proc boxplot data=x;
plot var1*n(trt time) / cboxfill=(c) haxis=axis1 vaxis=axis2;
format trt trt. time tm.;
Registration is open! SAS is returning to Vegas for an AI and analytics experience like no other! Whether you're an executive, manager, end user or SAS partner, SAS Innovate is designed for everyone on your team. Register for just $495 by 12/31/2023.
If you are interested in speaking, there is still time to submit a session idea. More details are posted on the website.
Learn how use the CAT functions in SAS to join values from multiple variables into a single value.
Find more tutorials on the SAS Users YouTube channel.