topic How to plot both scatterplot data and the fitted curve in Statistical Procedures

How to plot both scatterplot data and the fitted curve

Goyo — Thu, 18 Mar 2021 20:46:19 GMT

My first post to SAS communities...

I want the fitted data (red dots) in the scatterplot below to be a line rather than a set of dots. I don't see a way an option within the scatter statement to change the circles to lines. Any suggestions? (My code below.)

proc sgplot data=pred1;
scatter x=eucdist y= HTp; *blue circles;
scatter x=eucdist y= HTp_pred; *red circles;
run;

Re: How to plot both scatterplot data and the fitted curve

Reeza — Thu, 18 Mar 2021 21:08:09 GMT

SCATTER creates a scatter plot.
Use SERIES to create a line plot instead.
Everything else can stay the same.

EDIT: REG can do both at once though, without the pre-calculation as well. Remove the CLM/CLI to remove the prediction and confidence intervals. You could do all three together to test that if you'd like.

proc sgplot data=sashelp.class;
  reg x=height y=weight / CLM CLI;
run;

Re: How to plot both scatterplot data and the fitted curve

ballardw — Thu, 18 Mar 2021 21:31:56 GMT

@Reeza wrote:

SCATTER creates a scatter plot.
Use SERIES to create a line plot instead.
Everything else can stay the same.

EDIT: REG can do both at once though, without the pre-calculation as well. Remove the CLM/CLI to remove the prediction and confidence intervals. You could do all three together to test that if you'd like.
proc sgplot data=sashelp.class;
  reg x=height y=weight / CLM CLI;
run;

Data shows enough of a curve you may want the DEGREES=2 option to fit a quadratic in the Sgplot REG statement.

Re: How to plot both scatterplot data and the fitted curve

Rick_SAS — Mon, 22 Mar 2021 14:34:07 GMT

proc sort data=pred1; by eucdist; run;

proc sgplot data=pred1;
scatter x=eucdist y= HTp; *blue circles;
series x=eucdist y= HTp_pred / lineattrs=GraphData2; *line in contrasting color;
run;