topic Re: Grouped boxplots (with 3 grouping vars): Related to Sample 39101: Grouped box plot with droplin in New SAS User

Grouped boxplots (with 3 grouping vars): Related to Sample 39101: Grouped box plot with droplines

aebabinec — Wed, 17 Oct 2018 16:05:57 GMT

Hello,

I am using the code from here: http://support.sas.com/kb/39/101.html

However, I need help modifying it--

using the sample data, supposed I had another variable denoting health-clinic. I need to produce the box plot for each clinic, so how can I modify the code to produce 50 graphs instead of one overall?

On these graphs, I also need a to add a reference point for that health-clinics unique value. (on the baseline plot and 15month plot, by measure (see sample data))

I.E. the 50 graphs (just like the one in Sample 39101) will be the exact same except for the unique markers for each practice.

Thank you!

SAMPLE DATA:

PracticeID ClinicName Measure ProportionBL Proportion15

6001 hdosnonds aspirin 0.786 0.897

6002 hskndon aspirin 0.638 0.579

6003 hdonwod aspirin 0.526 0.599

6001 hdosnonds bloodp 0.986 0.857

6002 hskndon bloodp 0.238 0.903

6003 hdonwod bloodp 0.436 0.822

Re: Grouped boxplots (with 3 grouping vars): Related to Sample 39101: Grouped box plot with droplin

Reeza — Wed, 17 Oct 2018 16:50:15 GMT

That example is from 2010, SGPLOT now supports box plots, but not sure they're what you want.

An example is here:

https://support.sas.com/rnd/datavisualization/yourGraphs/analyticalQuick/vbox/

If you use SGPLOT then you can just add a BY statement for the CLINIC types to create a graph for each clinic.

Or try just adding a BY statement to the SGRENDER for clinics, It does seem to support BY group processing as well.

https://documentation.sas.com/?docsetId=grstatproc&docsetTarget=n131vj4eka12zpn1iwbqufke8l4z.htm&docsetVersion=9.4&locale=en

@aebabinec wrote:

Hello,

I am using the code from here: http://support.sas.com/kb/39/101.html

However, I need help modifying it--

using the sample data, supposed I had another variable denoting health-clinic. I need to produce the box plot for each clinic, so how can I modify the code to produce 50 graphs instead of one overall?

On these graphs, I also need a to add a reference point for that health-clinics unique value. (on the baseline plot and 15month plot, by measure (see sample data))

I.E. the 50 graphs (just like the one in Sample 39101) will be the exact same except for the unique markers for each practice.

Thank you!

SAMPLE DATA:

PracticeID ClinicName Measure ProportionBL Proportion15

6001 hdosnonds    aspirin 0.786 0.897

6002 hskndon aspirin 0.638    0.579

6003    hdonwod aspirin 0.526    0.599

6001 hdosnonds bloodp 0.986 0.857

6002 hskndon bloodp 0.238    0.903

6003    hdonwod bloodp 0.436    0.822

Re: Grouped boxplots (with 3 grouping vars): Related to Sample 39101: Grouped box plot with droplin

ballardw — Wed, 17 Oct 2018 16:57:31 GMT

It helps to show the code you are actually using. Referencing a source for similar code is nice but we can't tell what you may have implemented done that wasn't in the example.

By "unique marker" do you mean that you want 50 different markers such as different shape and color?

If you want 50 graphs, one per clinic then sort the data by clinic and add a by statement using the clinic variable.

Since the referenced code does not have a "baseline plot" or a "15month plot" your description is incomplete.

Re: Grouped boxplots (with 3 grouping vars): Related to Sample 39101: Grouped box plot with droplin

aebabinec — Wed, 17 Oct 2018 17:01:17 GMT

Here is the code-it is the same as in the sample, just changed to fit my data variables.

proc template;

define statgraph CQM_BY_Quarter;

begingraph;

entrytitle 'CQM Measures for &clinicname (Practice ID &PID) at Baseline and 15 Months';

*entryfootnote halign=left "For ALAT, ASAT and ALKPH, the Clinical Concern Level is 2 ULN" / textattrs=(size=8);

*entryfootnote halign=left "For BILTOT, the CCL is 1.5 ULN: where ULN is the Upper Level of Normal Range" / textattrs=(size=8);

layout overlay / cycleattrs=true yaxisopts=(label='Proportion Patients meeting tx guidelines')

xaxisopts=(display=(ticks tickvalues line));

boxplot x=measure y=proportionbl / discreteoffset=-0.2 boxwidth=0.2 display=(mean outliers caps fill)

outlineattrs=graphdata1(pattern=solid) meanattrs=graphdata1

medianattrs=graphdata1(pattern=solid) whiskerattrs=graphdata1(pattern=solid)

outlierattrs=graphdata1 name='a' legendlabel='Baseline';

boxplot x=measure y=proportion15 / discreteoffset= 0.2 boxwidth=0.2 display=(mean outliers caps fill)

outlineattrs=graphdata2(pattern=solid) meanattrs=graphdata2

medianattrs=graphdata2(pattern=solid) whiskerattrs=graphdata2(pattern=solid)

outlierattrs=graphdata2 name='b' legendlabel='15-Month';

discretelegend 'proportionbl' 'proportion15' / location=inside halign=right valign=top across=1;

*referenceline y=1 / lineattrs=(pattern=dot);

*referenceline y=1.5 / lineattrs=(pattern=dot);

*referenceline y=2 / lineattrs=(pattern=dot);

endlayout;

endgraph;

end;

run;

ods graphics / reset width=5in height=3in imagename= 'Fig1_1_CQM_by_Quarter';

ODS LISTING STYLE=LISTING IMAGE_DPI=100;

PROC SGRENDER DATA=WORK.TemplateData TEMPLATE=CQM_BY_QUARTER;

RUN;

For the unique markers- just markers for each practice for each measure on their respective baseline and 15month plots

Re: Grouped boxplots (with 3 grouping vars): Related to Sample 39101: Grouped box plot with droplin

aebabinec — Wed, 17 Oct 2018 21:06:18 GMT

@Reeza when I add the BY statement it only plots that clinics value (so I have one point for baseline and then one for 15 month)

any ideas? tahnk you for your help!

Re: Grouped boxplots (with 3 grouping vars): Related to Sample 39101: Grouped box plot with droplin

Reeza — Wed, 17 Oct 2018 21:59:36 GMT

Yes, that's what happens when you say you want '50 graphs', ie one for each clinic. If that's not what you want, you need to be more specific.