topic Re: JMP y vs x with "by" on the same plot in Discussions

JMP y vs x with "by" on the same plot

KurtAnderson — Fri, 09 Jun 2023 00:44:33 GMT

I have about 100 y (continuous) that I want to plot vs 1 x (continuous). The y-x plot is a nice fast way to do that. However, I also have a categorical variable that I would like to use as "by" (3 discrete values). If I use y-x and fill in the "by" field, I get 300 plots. I want 100 plots, one for each y, with the 3 "by's" plotted on the same plot. Alternatively if I use Graph Builder, I can only build one y vs x with "by" as the overlay. Then build the next, and next... Is there a fast way to make 100 y vs x that show all three "by" on each plot? Thanks.

Re: JMP y vs x with "by" on the same plot

Dan_Obermiller — Fri, 21 Jan 2022 16:26:29 GMT

Go to Fit Y by X, but do NOT specify a "By" variable. Make your graphs. Now, hold down the control key and go to the red triangle and select "Group By". Specify your categorical variable. Now, hold down the control key again and choose Fit Line.

The control key is acting as a broadcast command so that the menu choice applies to all 100 graphs.

Re: JMP y vs x with "by" on the same plot

SDF1 — Fri, 21 Jan 2022 16:36:11 GMT

Hi @KurtAnderson ,

Not sure if this is how you want to go about it, but you can do this in a quick way using the Column Switcher. If you generate your graph in Graph Builder and then go to the red hot button > Redo > Column Switcher, you'll want to select the current X column in the top window and then in the bottom window select all the other columns you want to switch it out with, including the current X variable. You'll get something like below (using the Car Physical Data.jmp file). Here, I'm using the column Type in the same way you're using your categorical variable. You can then click through the columns in column switcher and see each graph.

You can use the column switcher option in pretty much any platform within JMP. You might also want to play around with the role that you use your categorical variable, such as color, or overlay, or Group X, or Group Y to see which style best conveys the message you are trying to get across.

Hope this helps,

Re: JMP y vs x with "by" on the same plot

jthi — Fri, 21 Jan 2022 17:06:54 GMT

Besides using Column Switcher in Graph builder you could stack your Y variables and then use the label of Y-variables as Page and the data values as Y-axis.

Re: JMP y vs x with "by" on the same plot

KurtAnderson — Fri, 21 Jan 2022 18:02:17 GMT

Yes, Column Switcher works! Thanks.

Re: JMP y vs x with "by" on the same plot

KurtAnderson — Fri, 21 Jan 2022 18:05:21 GMT

Yes, Dan, that works, thank you. Initially I don't want all the stats so it's a little bit of an overkill, but the CTRL functionality is very nice. Thanks again.

Re: JMP y vs x with "by" on the same plot

KurtAnderson — Fri, 21 Jan 2022 18:06:50 GMT

Thanks JTHI. Funny, I do this with other data sets so why didn't I think of it? Thank you.

Re: JMP y vs x with "by" on the same plot

KurtAnderson — Fri, 21 Jan 2022 18:10:10 GMT

Too bad this functionality doesn't automatically change the markers to different shapes. The linear fit lines are different colors which is nice, but all the data markers default to black dots so I can't tell which dots belong to which linear fit line. Struggling a bit with color blindness, I'm a big fan of changing marker shape and color.

Re: JMP y vs x with "by" on the same plot

Mark_Bailey — Fri, 21 Jan 2022 19:39:41 GMT

You might use a row state (Color or Marker) to visually separate the levels of the categorical variable.

The Bivariate platform also has a command in the red triangle menu called Group By. These groups then separate the data internally if you use another command like Fit Line.