Box plots from five number summary - missing table?

[#1]

I am trying to understand the proposed solution (by @mpb ) here 

https://community.jmp.com/t5/Discussions/Box-plots-from-five-number-summary/td-p/2786

but the table suggested in the solution is missing. Can anyone help me with a similar table?

[#2]

Hi @Neo ,

apparently there are bits and pieces missing. What is it that you are trying to resolve?

Best

Florian

[#3]

@Florian_Vogt  Just trying to understand what the suggested in the solution (by @mpb ) on how to create a box plot when I have Min, Q1 (25%), Median, Q3 (75%) and Max values already calculated for each measurement on multiple parts. My data has three columns partID, Calculated 5 point summary (in the above order) and  Measured Parameter Name.

I am wondering if I can use the suggested solution in my case, but the table is missing in that solution for some reason.

[#4]

One solution might be this:

- if you have the 5 number summary, create an additional column with the frequencies

- in Graph Builder, create the boxplot and add the frequency column on the frequency drop zone in the bottom right corner

Table with Script is attached.

Does this help?

Best

Florian

[#5]

@Florian_Vogt  Thanks but unlikely to work in my case. Attached is my example data which is different from yours.

The 5 point summary values are ordered a bit differently than stated above, not sure if it matters. 

[#6]

I think you are missing some essential information.  You need to know how many observations you have (in your case, for each partID).  For example, if partID1 has 20 observations, then your Q1 data points occur with frequency .25*20 = 5.  So, if you have the total observations, you can construct the frequency column and follow the approach Florian provided.

[#7]

good point, my example is simple.

[#8]

I'll retract my comment.  I don't think you need the frequencies at all if you already have the Q1, Q3, etc. values.

[#9]

Hi @Neo,

sorry I submitted my answer without seeing your reply. Yes, the issue here is with the min and max I guess. If you were to use the 1% & 99% quantiles for example, it works fine with your data.

Best

Florian

[#10]

@dlehman1 @Florian_Vogt . Thanks but I think we have deviated from my original request. 

@mpb suggests the following. I would like to understand how a Fit Y by X could be used for example and what data form is necessary for that.