957632_en-US

liliconcom

How do I make a histogram of a percentage of failing value Y as a function of value X

Hi,

I have a set of continuous values Y for which I have a upper limit. I am interested in the percentage of failing values as a function of a second continuous value X. I was thinking of plotting this as a histogram, but I didn't find how to get JMP to create small intervals of my value X for the analysis.

Do someone have a solution for this kind of problem please?

Basic Data Analysis and ModelingRe: How do I make a histogram of a percentage of failing value Y as a function of value X

There might be a few ways to do it, but I would use a Capability Analysis (to define which points are above or below the upper limit), combined with a Logistic Fit (used to relate a Pass/Fail result to a continuous, numerical X variable). For example: 

 

1. Analyze -> Distribution, to create a histogram of "Y". 
2. Red Triangle -> Process Capability see what proportion of "Y" falls outside specification (above the upper limit) 
3. Red Triangle -> Save In Spec Indicator Formula to create a new column which equals 1 if "Y" is in spec, 0 if "Y" is out of spec. 
4. Analyze -> Fit Y by X using the newly created "Y in Spec" column as the Response, and "X" as the Factor. 

 

Here's how it looks when I follow the steps above, using the "Tablet Production.jmp" sample dataset: 

Edit: I realize I didn't address whether you can make a set of histograms. You can do this in graph builder using your "Y" variable in the Y-Axis Role, "X" in the Group X role, and the Histogram graph type. For example: 

 

Re: How do I make a histogram of a percentage of failing value Y as a function of value X

Thank you very much! The second solution is what I had in mind, but the first one gives a result with strong visual impact!