topic Re: Endpoints for Histogram (or remove outliers) in SAS Software for Learning Community

Endpoints for Histogram (or remove outliers)

chester2018 — Tue, 01 Nov 2022 19:43:24 GMT

Hi! I'm making a histogram of minutes of exercise in a week. Below is my graph and my code. I want to set up endpoints, so I can have a graph that makes sense. Or in other words, I want to have a histogram without the extreme observations

Re: Endpoints for Histogram (or remove outliers)

ballardw — Tue, 01 Nov 2022 22:12:11 GMT

In SGPLOT if your provide an XAXIS statement with a values list that will truncate values.

Such as:

xaxis values=(0 to 600 by 100);

which would limit the X values used to a maximum of 600.

Please provide code as TEXT pasted into a text box opened on the forum with the </> icon. I am not going to retype your code from a picture. It is hard to make changes to pictures.

Re: Endpoints for Histogram (or remove outliers)

chester2018 — Wed, 02 Nov 2022 04:18:16 GMT

Thanks for this suggestion, but the graph pretty much combined all the values into one big column. I also tried to make the upper limit 300. But then the output is zoom into the graph.

*this is for the first histogram*
proc sgplot data=aerobics_v5;
histogram Aerobic_Min_Week;
xaxis values=(0 to 1000 by 20);
run;

*this is for the second histogram*
proc sgplot data=aerobics_v5;
histogram Aerobic_Min_Week;
xaxis values=(0 to 300 by 20);
run;

Re: Endpoints for Histogram (or remove outliers)

chester2018 — Wed, 02 Nov 2022 04:21:22 GMT

I guess bin sizes would help

Re: Endpoints for Histogram (or remove outliers)

Rick_SAS — Wed, 02 Nov 2022 13:07:18 GMT

Read about how to use the BINWIDTH= option on the HISTOGRAM statement to define the bin width.

You can also use the MAX= option on the XAXIS statement to view only a subset of the histogram.

For example, the following statements truncate the range of the MPG_City variable in the Sashelp.Cars data and set the bin width to 2.

title "Zoom in on Lower Values of MPG_City";
title2 "Data Truncated by MPG <= 31";
proc sgplot data=Sashelp.cars;
   histogram MPG_City / binwidth=2 binstart=10;
   xaxis values=(10 to 30 by 2) max=31;  /* Optional: use VALUESHINT option */
run;