how do I create a probability plot of data across multiple lots

Question

I'm trying to create a probability plot of data, where three lots are plotted against each other.  Something like this:

However, I can't seem to get there from the data table.  I'm using JMP17.

hogi · Answer

"with some JSL magic" ...

Without JSL: use the Graph Builder Statistics, they adapt automatically to the changed settings.
Unfortunately - the chance to get Cumulative Probability included is VERY low.
IF CDF is fine, you can use:

Open( "$SAMPLE_DATA/Semiconductor Capability.jmp" );
New Column( "count",	Formula( 1 ));
Graph Builder(
	Variables( X( :INM2 ), Y( :count ), Overlay( :lot_id ) ),
	Elements(
		Points( X, Y, Legend( 3 ), Summary Statistic( "Cumulative Percent" ) ),
		Smoother( X, Y, Legend( 4 ), Summary Statistic( "Cumulative Percent" ) )
	),
	Column Switcher( :INM2, {Column Group( "Processes" )} )
);

EOwl · Answer

You've hinted at Column Switcher, is there is a way for a single column switcher to change the column (X-axis) and the transformed one (Y-Axis)? I've got about 40 parametric columns and I'd like to be able to toggle through them.

hogi · Answer

There are multiple possibilities available in JMP:
how to make cumulative probability plots in JMP?

Oneway [which is also used for your plot]  is by far the easiest approach.
Just activate the Normal Quantile plot via:

Life Distribution is  the most professional platform in Jmp to create such plots.
It allows to fit the data and to split the data via Grouping and the the By option.
Via JSL, one can collect the individual plots to generate a binned / grouped probability plot like this one:

If you don't care about the "-1" counting difference between cumulative percent and cumulative probability, you can use the summary statistics of Graph Builder.
There are many benefits:

* multiple parameters can be plotted next to each other
* compatible with all the functionality in Graph Builder, one can use Overlay, split the data into subplots.
* This approach is even compatible with Columns Switcher. This allows you to change the analysis parameter in the final report. Very helpful for exploratory data analysis!

The only disadvantage is the missing "N-1"

With the new Scoping Options of Graph Builder / JMP19, I will have a new favorite:

Then graphs like this one can be generated in a few seconds - and the user can drag columns onto the Overlay and Grouping dropzones to dig deeper and find hidden secrets.

And Column Switcher can be enabled with some JSL magic ...

Add CDFs to Graph Builder

Summary Statistics in Graph Builder - add Cumulative Probability