topic Re: Histogram Limited To Subset of Data? in SAS Procedures

Histogram Limited To Subset of Data?

NKormanik — Sat, 09 Jul 2016 07:52:20 GMT

I'd like to cut off really long tails from a histogram. There are a few outliers at either side. Histogram shows everything bunched up in the very middle, as a result.

SAS instructions are sorta crazy. A zillion options. But I can't find the option for this need.

I'd like to ONLY histogram data between, say, 100 and 300.

I could create a new dataset, as needed. But I'm hoping there is a more direct approach, by using some option on the histogram line, or in the univariate line. Or maybe some other ingenious way?

Thanks!

Nicholas Kormanik

Re: Histogram Limited To Subset of Data?

jklaverstijn — Sat, 09 Jul 2016 07:59:41 GMT

Usually you can control the number and values of your midpoints. If you'd like to share your data and your current attempt and results we may be better able to help you.

Regards, Jan.

Re: Histogram Limited To Subset of Data?

NKormanik — Sat, 09 Jul 2016 08:21:49 GMT

proc univariate data=have;
var N;
histogram;
run;

Data range from 1 to 4000. Mean=200. Total rows=A million.

The early numbers are bogus, as are the high numbers. I'd like to cut those off, and just see a histogram of the stuff in the middle.

Best would be to cut off via standard deviation, + or - one sd.

But it would be okay to use actual values, histogram values between 100 and 300.

Re: Histogram Limited To Subset of Data?

jklaverstijn — Sat, 09 Jul 2016 08:40:18 GMT

Easiest way out would be to add

where N between 100 and 300;

Hope this helps,

- Jan.

Re: Histogram Limited To Subset of Data?

NKormanik — Sat, 09 Jul 2016 09:47:10 GMT

Thanks Jan! Wow, was that easy, or what?! Beautiful!