How to 2X2 Chart ?!

[hellohere]

Say there is dataset contains X and 4Ys(Y1, Y2,Y3, and Y4).

How to generate a chart with 4-sub-charts (Y1~X, Y2~X, Y3~X and Y4~X)?!

[PaigeMiller]

Explain in more detail. Show us a sample of the input data. Show us a mocked up version of the desired output.

[hellohere]

One X with 4 Ys. One way is to plot 4Ys(below) with X within one chart.

When 4 Ys' Range varies and plot together does not make sense,

I like to make one chart with 4 carpets, each charting one Y with the same X

(SAS generate one Picture)

data temp;
do x=1 to 20;
y1=ranuni(1);y2=ranuni(2);y3=ranuni(3);y4=ranuni(4);
output;
end;
run;quit;

symbol1 interpol=join;symbol2 interpol=join;symbol3 interpol=join;symbol4 interpol=join;
proc gplot data=temp;
plot (y1 y2 y3 y4)*x/legend overlay;
run;quit;

[PaigeMiller]

First, your subject line is very misleading, there is no 2x2 charting in your question, and this confused me greatly. I suggest you go back to your original post and change it so that it has a more meaningful subject line.

Next, you provide data that really doesn't illustrate the problem. There's no different range of these 4 Y variables and they all fit nicely on one plot with one y-axis. As a recommendation to you, it is very helpful to provide data that illustrates your problem. It's never a good idea to make us guess.

Nevertheless at the risk of guessing wrong, here's some data that maybe closer to what you described, and a potential plot that might look better — but again, I am requesting that you tell us what you want instead of making us guess. Be specific, be detailed, be clear.

data temp;
do x=1 to 20;
y1=ranuni(1);
y2=10*ranuni(2);
y3=100*ranuni(3);
y4=1000*ranuni(4);
output;
end;
run;

proc sgplot data=temp;
series x=x y=y1;
series x=x y=y2;
series x=x y=y3;
series x=x y=y4;
yaxis type=log logstyle=logexpand logbase=10;
run;

[hellohere]

What wanted is stich the 4 plots into one(one pic/jpg), so that 4 takes four corners with the

final chart(top_left, top_right, bottom_left and bottom_right).

proc gplot data=sashelp.applianc(where=(cycle<=20));
plot units_1*cycle/legend overlay;
run;quit;
proc gplot data=sashelp.applianc(where=(cycle<=20));
plot units_2*cycle/legend overlay;
run;quit;
proc gplot data=sashelp.applianc(where=(cycle<=20));
plot units_3*cycle/legend overlay;
run;quit;
proc gplot data=sashelp.applianc(where=(cycle<=20));
plot units_4*cycle/legend overlay;
run;quit;

[PaigeMiller · Accepted solution]

ods layout gridded columns=2 rows=2 advance=proc;
proc sgplot data=sashelp.applianc(where=(cycle<=20));
scatter y=units_1 x=cycle;
run;
proc sgplot data=sashelp.applianc(where=(cycle<=20));
scatter y=units_2 x=cycle;
run;
proc sgplot data=sashelp.applianc(where=(cycle<=20));
scatter y=units_3 x=cycle;
run;
proc sgplot data=sashelp.applianc(where=(cycle<=20));
scatter y=units_4 x=cycle;
run;
ods layout end;

[hellohere]

any way to enforce the size of the outcome picture?! THX
See, stitch this way, 4 picture together, you still need scroll down-up-left-right to see all details.
If just fit the SAS OUTPUR window and no need to scroll, that would be perfect.

[PaigeMiller]

ods graphics/height=3in width=4in;

You can play with the width and height to find the numbers that look best in your situation.

[hellohere]

Thanks, It almost saved my day. BUT I tried below with varied inch #, the outcome picture
is same size... any help?


data temp;
do x=1 to 20;
y1=ranuni(1);
y2=10*ranuni(2);
y3=100*ranuni(3);
y4=1000*ranuni(4);
output;
end;
run;

ods layout gridded columns=2 rows=2 advance=proc;
ods graphics/height=1in width=2in;
proc gplot data=temp; plot y1*x; run;
proc gplot data=temp; plot y2*x; run;
proc gplot data=temp; plot y3*x; run;
proc gplot data=temp; plot y4*x; run;
ods layout end;

[PaigeMiller]

You need to use PROC SGPLOT and the proper syntax in the SCATTER or SERIES command.

This will not work with PROC GPLOT.

[ballardw]

---

@PaigeMiller wrote:

You need to use PROC SGPLOT and the proper syntax in the SCATTER or SERIES command.

This will not work with PROC GPLOT.

---

At least not without the headaches of Proc Template to create the display template and Proc GREPLAY to associate each already create plot with a designated space in the template.

[PaigeMiller]

PROC TEMPLATE + PROC GREPLAY is needed to replace ODS LAYOUT GRIDDED. I don't think that was clear from what you wrote.

The simplicity of ODS LAYOUT GRIDDED seems to win hands down here.

[hellohere]

Any way to save the picture with specified filename and file location, such as c:\mypic\_pic_today_01.jpg/png ?! Thanks,

[Ksharp]

data temp;
do x=1 to 20;
y1=ranuni(1);
y2=10*ranuni(2);
y3=100*ranuni(3);
y4=1000*ranuni(4);
output;
end;
run;

proc sgscatter data=temp;
plot (y1-y4)*x;
run;

[hellohere]

This is the one, one picture. Thanks,