import warnings
warnings.filterwarnings('ignore')
import pandas as pd
import numpy as np
import openml
from scipy.stats import chi2_contingency
import matplotlib.pyplot as plt
import seaborn as sns

dataset = openml.datasets.get_dataset(42178)
df, _, _, _ = dataset.get_data(dataset_format='dataframe')
print(df.shape)
print(df[['gender', 'SeniorCitizen', 'Churn']].head())

ct = pd.crosstab(df['gender'], df['Churn'])
print(ct)

chi2, p, dof, ex = chi2_contingency(ct)
print(f"Chi2 Statistic: {chi2:.3f}, p-value: {p:.4f}")

sns.countplot(x='gender', hue='Churn', data=df)
plt.title('Churn by Gender')
plt.show()

dataset = openml.datasets.get_dataset(1461)
df_marketing, _, _, _ = dataset.get_data(dataset_format='dataframe')
df_marketing.columns = ['age','job','marital','education','default','balance','housing','loan','contact','day','month','duration','campaign','pdays','previous','poutcome','response']
print(df_marketing[['job','response']].head())

ct_job = pd.crosstab(df_marketing['job'], df_marketing['response'])
print(ct_job.head())

chi2, p, dof, ex = chi2_contingency(ct_job)
print(f"Chi2 Statistic: {chi2:.2f}, p-value: {p:.5f}")
if p < 0.05:
    print('There is a statistically significant relationship between job type and campaign response.')
else:
    print('No significant relationship between job type and campaign response.')

ct_edu = pd.crosstab(df_marketing['education'], df_marketing['response'])
chi2, p, dof, ex = chi2_contingency(ct_edu)
print(ct_edu)
print(f"Chi2: {chi2:.2f}, p-value: {p:.4f}")

np.random.seed(42)
df_nps = pd.DataFrame({'CustomerID': range(1,501),
                      'Age': np.random.randint(18,70,500),
                      'Region': np.random.choice(['North','South','East','West'],500),
                      'NPS_Score': np.random.randint(0,11,500)})
df_nps['NPS_Category'] = pd.cut(df_nps['NPS_Score'], bins=[-1,6,8,10], labels=['Detractor','Passive','Promoter'])
print(df_nps.head())

ct_nps = pd.crosstab(df_nps['Region'], df_nps['NPS_Category'])
print(ct_nps)
chi2, p, dof, ex = chi2_contingency(ct_nps)
print(f"Chi2: {chi2:.2f}, p-value: {p:.4f}")

sns.countplot(x='Region', hue='NPS_Category', data=df_nps)
plt.title('NPS Category Distribution by Region')
plt.show()

age_bins = [17,29,39,49,59,70]
labels = ['18-29','30-39','40-49','50-59','60-69']
df_nps['AgeGroup'] = pd.cut(df_nps['Age'], bins=age_bins, labels=labels)
print(df_nps[['Age','AgeGroup']].head())

ct_age_nps = pd.crosstab(df_nps['AgeGroup'], df_nps['NPS_Category'])
print(ct_age_nps)
chi2, p, dof, ex = chi2_contingency(ct_age_nps)
print(f"Chi2: {chi2:.2f}, p-value: {p:.4f}")

ct_complex = pd.crosstab([df_nps['Region'], df_nps['AgeGroup']], df_nps['NPS_Category'])
print(ct_complex.head())

ct_stack = pd.crosstab(df_nps['Region'], df_nps['NPS_Category'], normalize='index')
ct_stack.plot(kind='bar', stacked=True, colormap='viridis')
plt.title('NPS Composition by Region (Proportion)')
plt.ylabel('Proportion')
plt.show()

df_nps_missing = df_nps.copy()
df_nps_missing.loc[::20, 'NPS_Category'] = np.nan
print(df_nps_missing['NPS_Category'].isnull().sum())
ct_missing = pd.crosstab(df_nps_missing['Region'], df_nps_missing['NPS_Category'])
print(ct_missing)

df_filtered = df_nps_missing.dropna(subset=['NPS_Category'])
ct_filled = pd.crosstab(df_filtered['Region'], df_filtered['NPS_Category'])
chi2, p, dof, ex = chi2_contingency(ct_filled)
print(f"After dropping missing: Chi2={chi2:.2f}, p-value={p:.4f}")

try:
    pd.crosstab(df_nps['Age'], df_nps['Region'])
    print('This produces a huge crosstab with too many groups for age.')
except Exception as e:
    print(str(e))

try:
    ct_wrong = pd.crosstab(df_nps['Region'], df_nps['NPS_Score'])
    chi2, p, dof, ex = chi2_contingency(ct_wrong)
    print(f"Chi2: {chi2:.2f}, p-value: {p:.4f}")
except Exception as e:
    print('Error:', e)

segments = df_nps.groupby('Region')['NPS_Score'].mean().sort_values()
print(segments)

ct = pd.crosstab(df_nps['Region'], df_nps['NPS_Category'], margins=True)
print(ct)

nps_summary = df_nps.groupby('Region')['NPS_Category'].value_counts(normalize=True).unstack().fillna(0)
nps_summary['NPS_Index'] = (nps_summary['Promoter'] - nps_summary['Detractor'])*100
print(nps_summary[['Promoter','Detractor','NPS_Index']])

# Trend analysis is not applicable here since survey responses are cross-sectional.
print('This dataset does not have a time column for trend analysis!')

ct = pd.crosstab(df['SeniorCitizen'], df['Churn'])
chi2, p, dof, ex = chi2_contingency(ct)
sns.countplot(x='SeniorCitizen', hue='Churn', data=df)
plt.title('Churn by Senior Citizen Status')
plt.show()
print(f"Chi2: {chi2:.2f}, p-value: {p:.4f}")
if p < 0.05:
    print('Recommendation: Offer special retention programs for senior citizens, as churn is significantly related.')
else:
    print('No specific retention strategy needed for seniors.')