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
michaelh
Level I

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. 

 

michaelh_0-1613332967229.png

 

1 ACCEPTED SOLUTION

Accepted Solutions
julian
Community Manager Community Manager

Re: Heatmap binning in JMP 15

Hi @michaelh,

I see just what you're experiencing -- setting the axis with an Increment=0.2, and Minor Ticks=0 returns this graph for me as well:

 

julian_3-1613510117845.png

 

Which, as you observed, isn't an increment of 0.2, but rather of 1. I think what's happening is the JMP axis display logic, to not have labels that overlap, is clashing with your specification. We can get around this by specifying an increment that JMP can display with the labels, such as an increment of 2, but then have the number of minor ticks specified to make intervals of 0.2 -- so, that's 9 minor ticks (at 0.2,0.4,0.6,0.8,1.0,1.2,1.4,1.6,1.8). Doing that returns:

 

julian_1-1613509955970.png

Does that work for you as a solution?

 

 

I don't recommend it, but one alternative to not have the axis display logic clash with the increment is to make the graph as big as possible, and decrease your font size until it's possible to display the labels for your increment. If you do that, nothing clashes and jmp will go ahead with an increment = 0.2 and no minor ticks.

 

julian_4-1613510298054.png

 

One final option, which I suggest only if you have every increment represented in the data, is to make the X axis column Nominal. That will force the display of each measured bin, with jmp packing the (now categorical) labeling the best it can:

nom-gb.gif

 

 

@julian 

 

 

 

 

View solution in original post

6 REPLIES 6
julian
Community Manager Community Manager

Re: 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 Nominal 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 

 

 

michaelh
Level I

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. 

julian
Community Manager Community Manager

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:

julian_0-1613418669810.png

 

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:

julian_1-1613418705521.png

 

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

julian_2-1613418791819.png

 

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

julian_3-1613418814512.png

 

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 

michaelh
Level I

Re: Heatmap binning in JMP 15

Thanks for sticking with this Julian...happy to share a data file. I've left a script for producing the heatmap in the table. 

As you can see, the script specificies an increment of 0.2, but produces a heatmap with increment of 1. For me this problem is specific to JMP 15 Pro (works fine in JMP 13). 

Also, hovering over a point in the heatmap confirms that each is an average of 5 rows. 

 

michaelh_0-1613508903967.png

 

 

 

julian
Community Manager Community Manager

Re: Heatmap binning in JMP 15

Hi @michaelh,

I see just what you're experiencing -- setting the axis with an Increment=0.2, and Minor Ticks=0 returns this graph for me as well:

 

julian_3-1613510117845.png

 

Which, as you observed, isn't an increment of 0.2, but rather of 1. I think what's happening is the JMP axis display logic, to not have labels that overlap, is clashing with your specification. We can get around this by specifying an increment that JMP can display with the labels, such as an increment of 2, but then have the number of minor ticks specified to make intervals of 0.2 -- so, that's 9 minor ticks (at 0.2,0.4,0.6,0.8,1.0,1.2,1.4,1.6,1.8). Doing that returns:

 

julian_1-1613509955970.png

Does that work for you as a solution?

 

 

I don't recommend it, but one alternative to not have the axis display logic clash with the increment is to make the graph as big as possible, and decrease your font size until it's possible to display the labels for your increment. If you do that, nothing clashes and jmp will go ahead with an increment = 0.2 and no minor ticks.

 

julian_4-1613510298054.png

 

One final option, which I suggest only if you have every increment represented in the data, is to make the X axis column Nominal. That will force the display of each measured bin, with jmp packing the (now categorical) labeling the best it can:

nom-gb.gif

 

 

@julian 

 

 

 

 

michaelh
Level I

Re: Heatmap binning in JMP 15

Thanks...Those workarounds work!