Exact mathematical formulas for SKEW() and KURTOSIS() chart functions

[martinmich]

Hello experts.

There could be different mathematical formulas for SKEW() and KURTOSIS() chart functions, e.g. just 3 different definitions are mentioned 1.3.5.11. Measures of Skewness and Kurtosis (nist.gov)

We would like to know exact formula implemented by Qlik for script function and also for the chart function. 

Unfortunately, For Kurtosis and Skew function official documentation does not mention exact implementation:
https://help.qlik.com/en-US/sense/May2024/Subsystems/Hub/Content/Sense_Hub/ChartFunctions/Statistica...

https://help.qlik.com/en-US/sense/May2024/Subsystems/Hub/Content/Sense_Hub/ChartFunctions/Statistica...

Can Qlik developers check and provide some clarity on this? 

Many thanks ahead. Regards, Martin 

[igoralcantara]

I do not know any documentation where this is disclaimed, but this is an interesting topic of investigation. Let’s wait to see if anyone here knows. If not, maybe I could do a little investigation, comparing the different parameters in R and see which one matches the results presented by Qlik.

[martinmich]

Yes it would be indeed great if Developers can provide the formula which is specified in the code but if this is not possible, we will probably have to do some "reverse engineering".

[igoralcantara]

Ok, as soon as I find some time I will work on it.

[igoralcantara]

Good news. I believe I will have that video published by Monday. Stay tuned.

[igoralcantara · Accepted solution]

I created a video answering your question at Data Voyagers. Please check here:

https://www.datavoyagers.net/post/the-math-behind-skewness-and-kurtosis-in-qlik

[martinmich]

Hello, Igoral. 

Thanks! All this looks very explanatory, and the formulas are there, so it is explicitly clear. In other words, now we know how Qlik processes data to obtain results for SKEW() and KURTOSIS().

Just one question: Is it possible to use weights within the formulas for the data? 

Regards, Martin