cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
Check out the JMP® Marketplace featured Capability Explorer add-in
Choose Language Hide Translation Bar
Edwinchiang
Level I

JMP heatmap Color display wrongly with nominal data

Hi

JMP seems to have a bug to show multi colors in one cell.

When I select "other" group, all green cell (other group) should be marked with dark color, but the cell I circled was not.

When I select the cell I circled previously, it is supposed to mark other group in legend. But it marked "edge" and "void" groups.

Any one knows how to display this multi color correctly, thank you in advance. 

 

Edwinchiang_2-1686310595214.png

Edwinchiang_3-1686310599956.png

 

 

1 ACCEPTED SOLUTION

Accepted Solutions

Re: JMP heatmap Color display wrongly with nominal data

I think what's happening is that the Mean of the rows represented by the cells is the Mean of Type's ordering index. With the new order, the Mean for that heatmap cell's Type is now between 0(edge) and 1(void), so both patterns are displayed, whereas before the Mean of Types 0(edge) and 2(void), was 1 (other). 

 

Before the reordering, when clicking on the other group, only cells representing rows with "other" as their Type were selected not cells with the Mean of the Types being "other".

 

When clicking on the cell that contained rows with "void" and "edge", only those two Types were highlighted in the legend.    

View solution in original post

2 REPLIES 2

Re: JMP heatmap Color display wrongly with nominal data

Can't say I'm 100% sure what the underlying problem is, but changing the value order of the Type column seems to fix the issue. Right click on the Type column header and go to Column Properties > Value Order, then set a custom order in which Other is either first or last. That yields blue-red crosshatching when Type is both Edge and Void.

Ross Metusalem
JMP Academic Ambassador

Re: JMP heatmap Color display wrongly with nominal data

I think what's happening is that the Mean of the rows represented by the cells is the Mean of Type's ordering index. With the new order, the Mean for that heatmap cell's Type is now between 0(edge) and 1(void), so both patterns are displayed, whereas before the Mean of Types 0(edge) and 2(void), was 1 (other). 

 

Before the reordering, when clicking on the other group, only cells representing rows with "other" as their Type were selected not cells with the Mean of the Types being "other".

 

When clicking on the cell that contained rows with "void" and "edge", only those two Types were highlighted in the legend.