In python, function to run tukey test on multiple csv files and export output in pdf format

Question

I have 20 csv files with each file having 4 columns state, dept, phase and budget.

I want to write a python function that does the following

1. Run tukey tests on each csv

2. Export the results as pdf files (2 per page if possible)

See below for 2 of the files (copy and paste the code to generate 2/20 of the csv files)

# The datasets generated below represents 2 out of the 20 files/datasets
# The 20 csv files have the same columns- state, dept, phase, budget
# 
import pandas as pd
import numpy as np
# genrating csv file 1 

df1={

'state':["PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA","PA"],
    
'dept':["hlth",'hlth','hlth','hlth','edu','edu','edu','edu','fin','fin','fin','fin','parks','parks','parks',"parks",'trans','trans','trans','trans'],
'phase':["one","one","one","one","one","one","one","one","one","one","one","one","one","one","one","one","one","one","one","one"],
'budget':[112.9,123,124,523.5,112,334,55,449,221.6,332,235,239,235,223,235.6,204,315.5,614,512,514.2]}

df1 = pd.DataFrame(df1)
df1


#generating csv file 2

df2={

'state':["NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY","NY"],
    
'dept':["hlth",'hlth','hlth','hlth','edu','edu','edu','edu','fin','fin','fin','fin','parks','parks','parks',"parks",'trans','trans','trans','trans'],
'phase':["two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two","two"],
'budget':[1212.9,1253,1244.4,5123.5,1312,3134,515.8,2449.9,3221.6,3132.5,2235.09,2239.01,3235.01,5223.01,4235.6,2204.5,2315.5,6114,4512,3514.2]}

df2 = pd.DataFrame(df2)
df2
# export the files to the folder

df1.to_csv(r'Path where you want to store the exported CSV file\PA_one.csv', index = False)
df2.to_csv(r'Path where you want to store the exported CSV file\NY_two.csv', index = False)

See below for my attempt

#libraries

from scipy import stats # for 1-way anova
from statsmodels.stats.multicomp import pairwise_tukeyhsd # for tukey 
#import seaborn as sns # for visualization
#import matplotlib.pyplot as plt # for visualization

import matplotlib.backends.backend_pdf
import glob # iterate through files in a folder
import os

path = "full\\path\\to_folder\\"  # 20 csv files
    
files = glob.glob(path+"\*.csv")

def tukey_reports(df,nFigPerPage):
    for file in files:    
        df = pd.read_csv(file)
        phase = df['phase'].drop_duplicates().sort_values()
        state = df['state'].drop_duplicates().sort_values()
      # set the pdf
        pdf = matplotlib.backends.backend_pdf.PdfPages("TukeyReports.pdf")
        nFig = 0
        iFigPerPage = 0
# 1-way anova 
        pval = stats.f_oneway(*[group["budget"].values for name, group in df.groupby("dept")])
        print(pval)
        if pval[1] < 0.05:
              print("*****samples are statistically different*****")
        else:
              print("*****samples are NOT statistically different*****")
# tukey test             
       tukey = pairwise_tukeyhsd(endog=df['budget'],groups=df['dept'],alpha=0.05)
       print(tukey)
       title = "state: {}, phase: {}".format(state, phase)
        
       iFigPerPage += 1
       if iFigPerPage % nFigPerPage == 0:
             iFigPerPage = 0

       nFig += 1
            
       if nFig % nFigPerPage == 0:
            fig.tight_layout()
            pdf.savefig()
               
                
    pdf.close()


# driver code
tukey_reports(df,nFigPerPage = 2)
    

I am looking for something like the figure below (2 figure per page)

Please share and comment your code for easier comprehension. Thanks in advance.