Normally, when you draw a histogram using JMP's Distribution Platform, the bin width is automatically set and drawn according to JMP's internal standards (data binning, number of observations).
The most common way to adjust the bin width of a JMP histogram is to use the grabber tool, but if the analyst wants to specify the bin width according to a specific rule ( Freedman–Diaconis rule ), the Set Bin Width option can be used.
The values required in the above formula are the IQR (Interquartile Range) of the variable to be analyzed and the sample size (n).
Let's take Diabetes.jmp, which is JMP Sample Data, as an example.
If a separate bin width is not specified, the histogram of the two continuous variables, BMI and BP, is expressed as follows.
For reference, the bin widths in the graph above are as follows:
bin width(BMI)=5, bin width(BP)=10
Freedman–Diaconis rule
If we calculate using the Freedman–Diaconis rule from the above data,
IQR(BMI)=6.15, IQR(BP)=21, n=442
bin width(BMI, Freedman–Diaconis rule)=1.614721725
bin width(BP, Freedman–Diaconis rule)=5.513683937
Therefore, if we apply the generated bin width, the graph above will change as follows.
Freedman–Deaconis JMP Add-In
Rather than calculating and applying the bin width manually as above, we created an add-in that can be applied using a simple JSL.
Below is a use case, a simple add-in that was created to avoid manual work on Freedman–Diaconis bin width. Also, the histogram defaults to stacked form.
If you need a histogram using the Freedman–Diaconis rule, I recommend installing and using it.
