358774_en-US

liliconcom

Heatmap binning in JMP 15

I am having an issue with heatmap binning in JMP Pro 15. My axis variable (time) is continuous and consists of 0.2s intervals. However, JMP 15 treats all increments <=1 exactly the same, plotting the heatmap with an increment of 1. 

 

This was not a problem in JMP13....the 2 plots below are the same data, same script.

 

Thanks in advance for any thoughts on what's going on here. 

 

 

GraphicsRe: Heatmap binning in JMP 15

Hi @michaelh,

I answered a question similar to yours in this post: 
Some problem using Heatmap.

 

In short, JMP 15 is attempting to keep the number of bins for the Heat Map to a reasonable number when you have a continuous modeling type. If you have summarized data at predictable intervals, you could treat that variable as Ordinal and JMP will plot every increment. Alternatively, as I show in the post I linked to above, you could instead go to the X Axis settings (double click the x-axis) and change the increment to match your desired level (0.2 in your case). 

 

I hope this helps!

@julian 

 

 

Re: Heatmap binning in JMP 15

Thanks Julian... I've tried both of those. Setting the incrememt to 0.2  for some reason still produces a heatmap with a 1.0 increment in JMP 15 -- the only difference between the 2 heatmaps I posted are the JMP version. The increments and all other axis/plot setting are exactly the same. Time is defined as continuous, but changing it to ordinal has no effect. 

Re: Heatmap binning in JMP 15

Hi @michaelh

 

I can't seem to replicate what you are experiencing in JMP 15: when I change the X-axis increment the Heat Map updates to that new setting. Let's see if we can determine if this is something specific with your data, or if there is something specific with your install of JMP going wrong.

 

Using the Iris.jmp sample data set, we can make the following heatmap, allowing for default scaling of the bins:

 

Open("$SAMPLE_DATA\Iris.jmp");
Graph Builder(
 Show Control Panel( 0 ),
 Variables( X( :Petal width ), Y( :Sepal length ) ),
 Elements( Heatmap( X, Y, Legend( 7 ) ) )
);

 

This will return:

 

Double-clicking the X-axis, we can see the Increment and minor ticks are set such that we will get a bin on the x at every 0.25 -- the increment of 0.5 divided by 1 minor tick into 2 sections:

 

If we wanted bins at every 0.10 increment, without minor ticks, we'd use:

 

and that returns this for me (after I make the window a little bigger to stop the crowing on the X axis):

 

Are you able to replicate this behavior interactively? If so, there must be something with your particular data table that's preventing the Heat Map from honoring your increment changes. Are you able to share a bit of your data?

 

@julian