topic Re: distribution without markers/dots in Graphics Programming

distribution without markers/dots

dustychair — Wed, 20 Nov 2019 19:23:14 GMT

Hi all,

I two numeric-Continuous variables. I'd like to draw a distribution where one continues variable x-axis and the other one is y-axis. I draw scatter plots with the curve but I don't want dots in the graph. It should be more like Kernel or Normal distribution. However, as I know I can draw Kernel or Normal distribution with only one variable using density. I want a distribution as seen below.

Could you help me to draw it?

Thanks.

Re: distribution without markers/dots

PeterClemmensen — Wed, 20 Nov 2019 19:24:16 GMT

Can you post your data or a sample of it?

Re: distribution without markers/dots

dustychair — Wed, 20 Nov 2019 19:27:48 GMT

Sure, Here it is! It is in the attachment.

Thanks

Re: distribution without markers/dots

dustychair — Wed, 20 Nov 2019 19:29:06 GMT

the scores will be x-axis and ss will be y-axis.

Thanks

Re: distribution without markers/dots

PeterClemmensen — Wed, 20 Nov 2019 19:34:36 GMT

I'm not really sure how you expect this data to yield a plot like you posted. However, to get things started, you can fit a normal density plot like this

data have;
input scores ss;
datalines;
920 5.5
941 6.3
935 8.2
910 9.5
972 10.5
954 .
925 9.2
900 8.7
;

proc sgplot data=have;
   density ss;
run;

Re: distribution without markers/dots

ballardw — Wed, 20 Nov 2019 21:40:10 GMT

Using @PeterClemmensen's data set.

Maybe:

proc sgplot data=have;
   pbspline x=scores y=ss;
run;

proc sgplot data=have;
   reg x=scores y=ss / degree=3 nomarkers;
run;

We kind of need a bit more description of your expectations than "no dots".

Re: distribution without markers/dots

dustychair — Thu, 21 Nov 2019 06:09:32 GMT

Thank you so much!

I tried the second one and it worked!

Thank you!