Is it possible to condtruct Mosaic Plots with equal widths for the X factors similar to what can be done with Box Plots?
Hi Smoore2,
Using graph builder you can create a mosaic plot with either proportionally sized or equally sized x categories.
If you use your X variable in the X drop zone, you will produce the typical mosaic plot with proportionally sized categories on the X
But, if you use your X variable in the "Group X" drop zone at the top, JMP will size the categories equally.
If you right click the category labels on the top, then select "X Group Edge" you can move the labels to the bottom (to most closely resemble the original mosaic plot).
I hope this helps!
Julian
Sorry for the typo: "Is it possible to construct...."
In a Mosaic Plot, the proportions on the x-axis represent the number of observations for each level of the X-variable. This is for explanatory value afforded by the plot to explain an additional dimension to the visualized graphic.
An alternative view that I like using graph builder is a heat map with the frequency color scale. This view uses equal real estate for the x-axis.
LouV, Yes, I understand why the columns are not the same width. Boxplots can do the same thing, but you can have the boxes the same width if desired. I was hoping this ability could be envoked for mosaic plots as well.
Thanks for your response.
smoore2,
Have you shared your desire with support@jmp.com so that it can be considered in the future?
I will do that. Thanks!
Hi Smoore2,
Using graph builder you can create a mosaic plot with either proportionally sized or equally sized x categories.
If you use your X variable in the X drop zone, you will produce the typical mosaic plot with proportionally sized categories on the X
But, if you use your X variable in the "Group X" drop zone at the top, JMP will size the categories equally.
If you right click the category labels on the top, then select "X Group Edge" you can move the labels to the bottom (to most closely resemble the original mosaic plot).
I hope this helps!
Julian
Thank-you, Julian!
This is a good solution for the simple situation as discussed.
If I want to generate a plot combining several variables, then I usually must put one variable in the X drop zone and then the proportional sizing is no longer available for this variable.
By the way, I usually am not free to select which variable is dropped in the x-zone because of the fixed priorization of x and Group x drop zones in the set-up of the plot.