[Python] In Python, How to run two statistical tests on all numeric columns

I have a dataframe df, I want to do the following:

1. run two stats tests on all the numeric columns (column_1 to column_84) to compare if there is a statistical difference between Types X, Y and Z

• 
  The stats tests are Kruskal and Dunn's tests
  
  
• 
  The comparing group: X vs Y, Y vs Z and X vs Z
  

1. Export the results to excel spreadsheet ( see screenshot below)

# copy & paste

## generate dataframe "df"

import pandas as pd
import numpy as np

df = pd.DataFrame(
data=np.random.uniform(low=5.5, high=30.75, size=(60, 84)),
columns=[f'column_{i}' for i in range(1, 85)],)

df.insert(loc=0, column='Type',value=np.repeat(['X','Y','Z'], 20, axis=0),)

df


I want to run kruskal wallis test and Dunn test for each column col_1 to col_84

# copy and paste the libraries below
from scipy.stats import kruskal
pip install scikit-posthocs
import scikit_posthocs as sp

# filtering for each Type X,Y and Z

# for column_1
# Extract values for each group
group_x_values = df[df['Type'] == 'X']['column_1'].values
group_y_values = df[df['Type'] == 'Y']['column_1'].values
group_z_values = df[df['Type'] == 'Z']['column_1'].values

# 1st stats test : Kruskal wallis

h_statistic, p_value = kruskal(group_x_values, group_y_values, group_z_values)

# Print the results
print(f"H-statistic: {h_statistic}")
print(f"P-value: {p_value}")


# 2nd stats test: Dunn test
data = [df[df['Group'] == 'X']['column_1'].values,
df[df['Group'] == 'Y']['column_1'].values,
df[df['Group'] == 'Z']['column_1'].values]

p_values = sp.posthoc_dunn(data, p_adjust='bonferroni')

print(p_values)

# for column_2

group_x_values = df[df['Type'] == 'X']['column_2'].values
group_y_values = df[df['Type'] == 'Y']['column_2'].values
group_z_values = df[df['Type'] == 'Z']['column_2'].values

# 1st stats test : Kruskal wallis
h_statistic, p_value = kruskal(group_x_values, group_y_values, group_z_values)
# Print the results
print(f"H-statistic: {h_statistic}")
print(f"P-value: {p_value}")

# 2nd stats test: Dunn test
data = [df[df['Group'] == 'X']['column_2'].values,
df[df['Group'] == 'Y']['column_2'].values,
df[df['Group'] == 'Z']['column_2'].values]
p_values = sp.posthoc_dunn(data, p_adjust='bonferroni')

print(p_values)
.
.
.
#for column_84
group_x_values = df[df['Type'] == 'X']['column_84'].values
group_y_values = df[df['Type'] == 'Y']['column_84'].values
group_z_values = df[df['Type'] == 'Z']['column_84'].values
# 1st stats test : Kruskal wallis
h_statistic, p_value = kruskal(group_x_values, group_y_values, group_z_values)
# Print the results
print(f"H-statistic: {h_statistic}")
print(f"P-value: {p_value}")

# 2nd stats test: Dunn test
data = [df[df['Group'] == 'X']['column_84'].values,
df[df['Group'] == 'Y']['column_84'].values,
df[df['Group'] == 'Z']['column_84'].values]

p_values = sp.posthoc_dunn(data, p_adjust='bonferroni')

print(p_values)



I want to export both results to excel worksheet, something like this:

Kruskal Worksheet



Dunn Worksheet



Continue reading...