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

dataset = openml.datasets.get_dataset(42178)
df, _, _, _ = dataset.get_data(dataset_format='dataframe')
print(df.shape)
print(df.head(3))

n_missing = df.isnull().sum().sum()
print(f"Total missing values in the survey dataset: {n_missing}")

churn_counts = df['Churn'].value_counts()
print("Customer Churn Counts:")
print(churn_counts)

gender_churn = df.groupby('gender')['Churn'].value_counts().unstack()
print(gender_churn)

churn_rate = df['Churn'].value_counts(normalize=True).loc['Yes']
print(f"Overall churn rate: {churn_rate:.2%}")

sns.countplot(data=df, x='Churn', hue='Contract')
plt.title('Churn by Contract Type')
plt.ylabel('Number of Customers')
plt.xlabel('Churn')
plt.show()

monthly_charges = df.groupby('Churn')['MonthlyCharges'].mean()
print("Average monthly charges by churn status:")
print(monthly_charges)

plt.figure(figsize=(8,4))
sns.boxplot(data=df, x='Churn', y='MonthlyCharges')
plt.title('Monthly Charges: Churned vs Not Churned Customers')
plt.show()

internet_churn = df.groupby('InternetService')['Churn'].value_counts(normalize=True).unstack()
internet_churn['Churn Rate (%)'] = internet_churn['Yes'] * 100
print(internet_churn[['Churn Rate (%)']])

survey_summary = df.describe(include='all')
survey_summary.to_csv('survey_summary.csv')
print('Survey summary statistics saved as survey_summary.csv')

campaign_dataset = openml.datasets.get_dataset(1461)
df_campaign, _, _, _ = campaign_dataset.get_data(dataset_format='dataframe')
df_campaign.columns = ['age','job','marital','education','default','balance','housing','loan','contact','day','month','duration','campaign','pdays','previous','poutcome','response']
pivot = pd.crosstab(df_campaign['job'], df_campaign['response'], normalize='index') * 100
pivot = pivot.rename(columns={'yes':'Yes', 'no':'No'}) if 'yes' in pivot.columns else pivot
print(pivot)

sns.heatmap(pivot, annot=True, fmt='.1f', cmap='coolwarm')
plt.title('Campaign Positive Response Rate by Job Segment (%)')
plt.ylabel('Job')
plt.xlabel('Response')
plt.show()

nps_df = 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)})
nps_df['NPS_Type'] = np.where(nps_df['NPS_Score'] >= 9, 'Promoter',
                    np.where(nps_df['NPS_Score'] <=6, 'Detractor', 'Passive'))
nps_summary = nps_df['NPS_Type'].value_counts(normalize=True) * 100
nps_score = nps_summary['Promoter'] - nps_summary['Detractor']
print('NPS Summary (%) by Type:')
print(nps_summary)
print(f'Overall Net Promoter Score (NPS): {nps_score:.1f}')

region_nps = nps_df.groupby('Region')['NPS_Type'].value_counts(normalize=True).unstack().fillna(0) * 100
region_nps['NPS'] = region_nps['Promoter'] - region_nps['Detractor']
print(region_nps[['Promoter','Passive','Detractor','NPS']])

missing_nps = nps_df['NPS_Score'].isnull().sum()
print(f"Number of missing NPS scores: {missing_nps}")

try:
    bad_group = nps_df.groupby('Region')['NPS_Scorez'].mean()
except Exception as e:
    print('Error grouping by wrong column name:', e)

if set(nps_df['NPS_Score'].unique()) - set(range(0,11)):
    print("Warning: Some NPS scores fall outside the 0-10 range!")
else:
    print("All NPS scores are within the valid 0-10 range.")

segmentation = df.groupby(['SeniorCitizen', 'Churn']).size().unstack()
print('Customer Segmentation by Senior Citizen Status and Churn:')
print(segmentation)

cross_tab = pd.crosstab(df['gender'], df['Contract'])
print('Cross-tabulation: Gender vs. Contract Type')
print(cross_tab)

df['Price_Category'] = pd.cut(df['MonthlyCharges'], bins=[0,40,70,150], labels=['Low','Mid','High'])
price_churn = df.groupby('Price_Category')['Churn'].value_counts(normalize=True).unstack().fillna(0) * 100
print(price_churn)

trend = df.groupby('tenure')['Churn'].value_counts(normalize=True).unstack().fillna(0)['Yes'] * 100
plt.figure(figsize=(8,4))
plt.plot(trend.index, trend.values)
plt.title('Churn Rate over Customer Tenure')
plt.xlabel('Tenure (months)')
plt.ylabel('Churn Rate (%)')
plt.show()

# 1. Load open-ended customer feedback
df_feedback = pd.DataFrame({'CustomerID':[1,2,3,4,5],
                           'Feedback':['Great service and friendly staff','Delivery was slow and packaging was poor','Excellent quality, will buy again','Customer support needs improvement','Good value for money']})
print(df_feedback)

# 2. Simple keyword extraction
keywords = pd.Series(' '.join(df_feedback['Feedback']).lower().split()).value_counts().head(5)
print('Most common feedback keywords:')
print(keywords)

# 3. Write an executive summary
summary = 'Key customer themes: Delivery speed and packaging need improvement, while service and quality are praised. Customers value support and good prices.'
with open('customer_feedback_summary.txt', 'w') as f:
    f.write(summary)
print('Summary saved as customer_feedback_summary.txt')