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

# Load customer satisfaction dataset (OpenML 42178)
dataset = openml.datasets.get_dataset(42178)
df, _, _, _ = dataset.get_data(dataset_format='dataframe')
print('Shape:', df.shape)
print(df.head(3))

# Simple bar chart: Gender distribution
gender_counts = df['gender'].value_counts()
plt.figure(figsize=(5,3))
sns.barplot(x=gender_counts.index, y=gender_counts.values, palette='pastel')
plt.title('Customer Gender Distribution')
plt.ylabel('Number of Customers')
plt.xlabel('Gender')
plt.tight_layout()
plt.show()

# Pie chart: Customer Churn Rate
churn_counts = df['Churn'].value_counts()
plt.figure(figsize=(4,4))
plt.pie(churn_counts, labels=churn_counts.index, autopct='%1.1f%%', colors=['#5ab4ac','#d8b365'])
plt.title('Customer Churn Rate')
plt.show()

# Countplot: Customer Contract Types
plt.figure(figsize=(6,3))
sns.countplot(x='Contract', data=df, palette='Set2')
plt.title('Distribution of Contract Types')
plt.xlabel('Contract Type')
plt.ylabel('Number of Customers')
plt.tight_layout()
plt.show()

# Histogram: Monthly Charges Distribution
plt.figure(figsize=(6,3))
plt.hist(df['MonthlyCharges'], bins=20, color='#4575b4', edgecolor='black')
plt.title('Monthly Charges Distribution')
plt.xlabel('Monthly Charges ($)')
plt.ylabel('Number of Customers')
plt.tight_layout()
plt.show()

# Boxplot: Monthly Charges by Churn Status
plt.figure(figsize=(6,4))
sns.boxplot(x='Churn', y='MonthlyCharges', data=df, palette=['#99d594','#fc8d62'])
plt.title('Monthly Charges by Churn Status')
plt.xlabel('Churn')
plt.ylabel('Monthly Charges ($)')
plt.tight_layout()
plt.show()

# Stacked Bar Chart: Internet Service Type by Churn
ct = pd.crosstab(df['InternetService'], df['Churn'])
ct.plot(kind='bar', stacked=True, color=['#5ab4ac','#d8b365'], figsize=(7,4))
plt.title('Internet Service Type by Churn')
plt.xlabel('Internet Service Type')
plt.ylabel('Number of Customers')
plt.xticks(rotation=0)
plt.legend(title='Churn')
plt.tight_layout()
plt.show()

# Trend Line: Average Monthly Charges by Tenure
avg_charges_by_tenure = df.groupby('tenure')['MonthlyCharges'].mean()
plt.figure(figsize=(8,3))
plt.plot(avg_charges_by_tenure.index, avg_charges_by_tenure.values, color='#fee08b', marker='o')
plt.title('Average Monthly Charges by Tenure (Months)')
plt.xlabel('Tenure (Months)')
plt.ylabel('Avg Monthly Charges')
plt.grid(True, linestyle=':', alpha=0.5)
plt.tight_layout()
plt.show()

# Setup: Generate synthetic NPS survey dataset
np.random.seed(42)
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)
})
print(nps_df.head(3))

# Bar chart: NPS Score Distribution
plt.figure(figsize=(7,3))
sns.countplot(x='NPS_Score', data=nps_df, palette='Blues')
plt.title('Customer NPS Score Distribution')
plt.xlabel('NPS Score (0=Detractor, 10=Promoter)')
plt.ylabel('Number of Responses')
plt.tight_layout()
plt.show()

# Grouped Bar: Mean NPS Score by Region
nps_region = nps_df.groupby('Region')['NPS_Score'].mean().reset_index()
plt.figure(figsize=(6,3))
sns.barplot(x='Region', y='NPS_Score', data=nps_region, palette='Set1')
plt.title('Average NPS Score by Region')
plt.ylabel('Average NPS Score')
plt.xlabel('Region')
plt.tight_layout()
plt.show()

# Setup: Sample open-ended customer feedback
feedback_df = 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(feedback_df)

# Wordcloud for customer feedback
from wordcloud import WordCloud
text = ' '.join(feedback_df['Feedback'])
wordcloud = WordCloud(background_color='white', width=400, height=200).generate(text)
plt.figure(figsize=(7,3))
plt.imshow(wordcloud, interpolation='bilinear')
plt.axis('off')
plt.title('Customer Feedback Wordcloud')
plt.show()

# Advanced: Cluster customers by tenure and monthly charges
from sklearn.preprocessing import StandardScaler
from sklearn.cluster import KMeans
clust_df = df[['tenure','MonthlyCharges']].dropna()
scaler = StandardScaler()
scaled = scaler.fit_transform(clust_df)
kmeans = KMeans(n_clusters=3, random_state=42)
labels = kmeans.fit_predict(scaled)
clust_df['Cluster'] = labels
plt.figure(figsize=(7,4))
sns.scatterplot(x='tenure', y='MonthlyCharges', hue='Cluster', data=clust_df, palette='viridis')
plt.title('Market Segmentation by Tenure & Charges')
plt.xlabel('Tenure (Months)')
plt.ylabel('Monthly Charges ($)')
plt.legend(title='Cluster')
plt.tight_layout()
plt.show()

# Generate synthetic cohort dataset
np.random.seed(0)
dates = pd.date_range('2021-01-01', periods=24, freq='ME')
cohort_df = pd.DataFrame({
    'CustomerID': np.random.randint(1000,2000,len(dates)),
    'Signup_Month': dates,
    'Active_Users': np.random.randint(50,300,len(dates))
})
print(cohort_df.head(3))

# Line plot: Cohort Retention Trend
plt.figure(figsize=(8,4))
plt.plot(cohort_df['Signup_Month'], cohort_df['Active_Users'], marker='s', color='#b2182b')
plt.title('Monthly Active Users by Cohort Signup Month')
plt.xlabel('Signup Month')
plt.ylabel('Active Users')
plt.grid(True, ls=':', alpha=0.4)
plt.tight_layout()
plt.show()

# Detect missing values in survey data
missing = df.isnull().sum()
print('Missing survey responses per column:')
print(missing[missing > 0])

# Visualize missing survey data
import missingno as msno
msno.matrix(df.sample(500, random_state=42))
plt.title('Missing Survey Data Visualization')
plt.show()

# Example: Incorrect aggregation warning
try:
    avg = df.groupby('Churn')['MonthlyCharges'].sum()['Yes'] / df['Churn'].value_counts()['Yes']
    print('Average monthly charge (churned):', avg)
except Exception as e:
    print('Aggregation error:', e)

# Example: NPS scale misinterpretation
promoters = (nps_df['NPS_Score'] >= 9).sum()
detractors = (nps_df['NPS_Score'] <= 6).sum()
nps_score = 100 * (promoters - detractors) / len(nps_df)
print('Net Promoter Score (NPS):', round(nps_score, 1))

# Best Practice: Cross-tabulation of Churn by Contract Type
ctab = pd.crosstab(df['Contract'], df['Churn'])
print(ctab)

# Build a satisfaction score index from available indicators
df['Satisfaction_Score'] = (df['tenure']/df['tenure'].max())*0.4 + (df['MonthlyCharges']/df['MonthlyCharges'].max())*0.6
print(df[['tenure','MonthlyCharges','Satisfaction_Score']].head(3))

# Setup: Load bank marketing campaign dataset
dataset2 = openml.datasets.get_dataset(1461)
df2, _, _, _ = dataset2.get_data(dataset_format='dataframe')
df2.columns = ['age','job','marital','education','default','balance','housing','loan','contact','day','month','duration','campaign','pdays','previous','poutcome','response']
month_summary = df2['month'].value_counts().sort_index()
plt.figure(figsize=(7,3))
sns.lineplot(x=month_summary.index, y=month_summary.values, marker='o', color='#5e4fa2')
plt.title('Campaign Calls by Month')
plt.xlabel('Month')
plt.ylabel('Number of Calls')
plt.grid(ls=':', alpha=0.4)
plt.tight_layout()
plt.show()

# Mini Challenge: Identify if senior citizens are more likely to churn
senior_table = pd.crosstab(df['SeniorCitizen'], df['Churn'])
print('Senior Citizen by Churn Table:')
print(senior_table)
senior_churn_rate = senior_table.loc[1,'Yes'] / senior_table.loc[1].sum()
non_senior_churn_rate = senior_table.loc[0,'Yes'] / senior_table.loc[0].sum()
print('Churn Rate (Senior):', round(senior_churn_rate*100,1),'%')
print('Churn Rate (Non-Senior):', round(non_senior_churn_rate*100,1),'%')

# Save current figure as a PDF for business reporting
fig, ax = plt.subplots(figsize=(6,3))
sns.countplot(x='Contract', data=df, palette='Set2', ax=ax)
ax.set_title('Distribution of Contract Types')
plt.tight_layout()
fig.savefig('contract_types_distribution.pdf')