Kertosis and Skewness

[SteveP]

Other than going to the raw math, anyone figure out how to obtain Kertosis and Skewness measures of a univariate distribution? Macro?

[jdunkerley79 · Accepted solution]

Easiest and quickest way I can think of is to use Python:

First get packages:

from ayx import Package
Package.installPackages(['pandas','scipy'])

Next:

from ayx import Alteryx
import scipy.stats
import pandas as pd
a = Alteryx.read('#1')
kurt = scipy.stats.kurtosis(a['Field1'])
skew = scipy.stats.skew(a['Field1'])
d = {'kurt': [kurt], 'skew': [skew]}
df = pd.DataFrame(data=d)
Alteryx.write(df,1)

Simple workflow using this attached.

[SteveP]

Good idea - now to figure out Python. I see a few Python help's in the community so I'll check them out. 

Thanks!

[marca_at_cue]

Can you help me modify the Python code to pass in another field to subgroup by when doing the kurtosis and skew calculations?  Output would be 1 row per subgroup_value, with 3 fields in each row: subgroup_value, kurtosis, skew.

[SteveP]

Super - worked flawlessly! Thanks!

[LipkeJK]

Has anyone experienced error messages when using a python tool in a macro?  I'm trying to calculate the Skew for each product.  The python script in this thread works when I run it by itself, but I'm running into a roadblock when applying it to a macro.  

Thanks!

Jon