Data visualization with SAS programming

Box Plots

Accepted Solution Solved
Reply
Occasional Contributor
Posts: 16
Accepted Solution

Box Plots

I have output my interactive means from the glimmix procedure of SAS and am trying to make box plots from these means. Please see the attached. I don't understand why they aren't coming through as box plots. See my code below. I am wanting to look and see if there are any outliers in this data set and thus, am wanting to visualize with box plots. Any suggestions?

 

Thanks!

 

Proc SGPlot data=Means;
Vbox Mu/ Category=ParityGrpStrat1 Group=Trt;
Run;

 

 Outliers.png


Accepted Solutions
Solution
‎12-29-2017 11:32 AM
SAS Super FREQ
Posts: 496

Re: Box Plots

Given your data, you are successfully plotting the means.  You see the mean and the median, which are the same.  That is all you provided to plot.

View solution in original post


All Replies
SAS Super FREQ
Posts: 496

Re: Box Plots

Try it once without group= and see if that is more what you had in mind.  I have never tried group= and category= together.  Do you have any variability within those groups?

Occasional Contributor
Posts: 16

Re: Box Plots

Posted in reply to WarrenKuhfeld

Thanks. This allowed me to see the box plots better for each "grouping" if you will. I plotted with the raw data and was successful. Just not sure why it wouldn't work the same with these means.

 

Thanks for your help!

SAS Super FREQ
Posts: 9,261

Re: Box Plots

Hi:

  It was my understanding that the Box plot was doing the summarizing for you. For example, in the documentation about box plots:

about_box_plots.png

SASHELP.PRDSALE is NOT pre-summarized. The VBOX statement causes the data to be summarized and the mean calculated.

 

Or, if I do this:

proc sgplot data=sashelp.iris;
  vbox sepallength / group=species;
run;

SASHELP.IRIS is not presummarized, the mean and outlier are shown based on the unsummarized data in the data set.

 

The things I might suggest are that you use your original dataset with the SGPLOT procedure step so that you see the distribution of your unsummarized data.

 

OR, PROC GLIMMIX has a PLOTS= option and will produce a Box plot using ODS GRAPHICS, so looking up the PLOTS= option might help you too.

 

Also, if you are just looking for outliers, PROC UNIVARIATE will show you the 5 highest and lowest (extreme obs) by default if you do this:

proc univariate data=sashelp.iris;
  var sepallength;
run;

cynthia

 

Occasional Contributor
Posts: 16

Re: Box Plots

Posted in reply to Cynthia_sas

Thanks for your help. I was able to successfully plot with the raw data. I am still not sure though why I can't with the means. 

 

Thanks again!

Solution
‎12-29-2017 11:32 AM
SAS Super FREQ
Posts: 496

Re: Box Plots

Given your data, you are successfully plotting the means.  You see the mean and the median, which are the same.  That is all you provided to plot.

☑ This topic is solved.

Need further help from the community? Please ask a new question.

Discussion stats
  • 5 replies
  • 185 views
  • 0 likes
  • 3 in conversation