Subscribe Bookmark RSS Feed

Auto-generated stacked histogram

JesperJohansen

Community Trekker

Joined:

Apr 15, 2013

Hi

I am looking for a function to generate a histogram with color information on data groups like this:

Graph Builder.png

However, in this case I needed to bin the data myself. Also the bars don't touch like normally in histograms.

 

I know the function in graph builder generating the type of histogram shown below, but here the istograms are not stacked but superimposed on each other, and almost impossible to read for many groups for many groups:

Graph Builder 2.png

I know that selectiung data in the Distribution platform also shows the selected data in the histogram. But here I can only show one group (color) at the time.

Is ther no built-in function in JMP for this?

 

I hope someone can help. I attached the data with scripts.

BR
Jesper
2 REPLIES
txnelson

Super User

Joined:

Jun 22, 2012

I am not aware of a method to get your results exactly the way you want, the below script is what I have used for years to get close to those results in an automated fashion.  I set it up to run on your attached data table, but it could be easily generalized.

stacked.PNG

Names Default To Here( 1 );
dt = Current Data Table();
dis = dt << Distribution(
	invisible,
	Continuous Distribution(
		Column( :x ),
		Outlier Box Plot( 0 ),
		Normal Quantile Plot( 1 ),
		Customize Summary Statistics(
			Skewness( 1 ),
			Kurtosis( 1 ),
			Median( 1 ),
			Median Absolute Deviation( 1 ),
			Robust Mean( 1 ),
			Robust Std Dev( 1 )
		)
	),
	Nominal Distribution( Column( :Group ) )
);
// Save the bins
dis << save( Level Midpoints );
// Change the X axis modeling type so the bins are equal distance groupings
dt:Midpoint x << modeling type( ordinal );
// I use the Chart Platform for this because I developed the method long
// before the Graph Builder was available
Chart( X( :Midpoint x, :Group ), Y( N( :x ) ), Stack Bars( 1 ), Bar Chart( 1 ) );
Jim
JesperJohansen

Community Trekker

Joined:

Apr 15, 2013

Than you Jim

I was not aware of the option to save midpoints from distribution. That of course is a neat way to bin the data.

I guess you answer is that the feature does not exist as a built-in function.

 

To JMP: Consider this a feature request then :)

BR
Jesper