# ===================================================================
# BONUS LESSON 7: SUSTAINABLE FINANCE AND ESG ANALYTICS
# ===================================================================

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns
from scipy.stats import pearsonr
from sklearn.preprocessing import MinMaxScaler
import warnings
warnings.filterwarnings('ignore')

# Set style
sns.set_style("whitegrid")
np.random.seed(42)

print("="*70)
print("SUSTAINABLE FINANCE AND ESG ANALYTICS")
print("="*70)

# ----------------------------------------------------------------
# PART A: GENERATE SYNTHETIC ESG DATA FOR COMPANIES
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART A: ESG Data for a Portfolio of Companies")
print("-"*60)

# Simulate 100 companies with ESG scores and financial metrics
n_companies = 100
sectors = ['Technology', 'Finance', 'Healthcare', 'Energy', 'Consumer', 'Utilities', 'Materials']

company_data = []
for i in range(n_companies):
    sector = np.random.choice(sectors)
    # ESG pillars (0-100 scale)
    e_score = np.random.normal(50, 20).clip(0, 100)
    s_score = np.random.normal(50, 20).clip(0, 100)
    g_score = np.random.normal(50, 20).clip(0, 100)
    
    # Adjust for sector biases
    if sector == 'Energy':
        e_score = np.random.normal(30, 15).clip(0, 100)
    elif sector == 'Technology':
        e_score = np.random.normal(60, 15).clip(0, 100)
        s_score = np.random.normal(55, 15).clip(0, 100)
    elif sector == 'Utilities':
        e_score = np.random.normal(55, 10).clip(0, 100)
    
    # Controversies (0 = none, 1 = minor, 2 = major)
    controversies = np.random.choice([0, 1, 2], p=[0.7, 0.2, 0.1])
    
    # ESG total (weighted average with penalties for controversies)
    esg_total = 0.4 * e_score + 0.3 * s_score + 0.3 * g_score
    if controversies == 1:
        esg_total -= 5
    elif controversies == 2:
        esg_total -= 15
    esg_total = esg_total.clip(0, 100)
    
    # Financial metrics
    revenue = np.random.gamma(5, 200).clip(50, 5000)  # $M
    net_income = revenue * np.random.uniform(0.02, 0.15)
    market_cap = revenue * np.random.uniform(0.5, 3.0)
    
    # Carbon emissions (Scope 1+2) in tonnes per $M revenue
    carbon_intensity = np.random.lognormal(2, 1).clip(0.1, 100)
    if sector == 'Energy':
        carbon_intensity *= 3
    elif sector == 'Utilities':
        carbon_intensity *= 2
    elif sector == 'Technology':
        carbon_intensity *= 0.5
    
    company_data.append({
        'Company': f'Company_{i+1}',
        'Sector': sector,
        'E_Score': e_score,
        'S_Score': s_score,
        'G_Score': g_score,
        'ESG_Total': esg_total,
        'Revenue': revenue,
        'Net_Income': net_income,
        'Market_Cap': market_cap,
        'Carbon_Intensity': carbon_intensity,
        'Controversies': controversies
    })

df_esg = pd.DataFrame(company_data)
print(f"Generated data for {len(df_esg)} companies.")
print(df_esg.head().round(2))

# ----------------------------------------------------------------
# PART B: ESG PORTFOLIO CONSTRUCTION
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART B: ESG Portfolio Construction")
print("-"*60)

# 1. Negative screening: Exclude companies with Controversies = 2 (major)
df_clean = df_esg[df_esg['Controversies'] < 2].copy()
print(f"After negative screening (exclude major controversies): {len(df_clean)} companies remain")

# 2. Positive screening: Select top ESG performers (top 30 by ESG_Total)
df_top_esg = df_clean.nlargest(30, 'ESG_Total')
print(f"Selected top 30 companies by ESG score.")

# 3. Create portfolio weights (equal-weighted)
df_top_esg['Weight'] = 1 / len(df_top_esg)

# 4. Sector allocation
sector_allocation = df_top_esg.groupby('Sector')['Weight'].sum() * 100

print("\nESG Portfolio Sector Allocation:")
print(sector_allocation.round(2))

# 5. Portfolio ESG score (weighted average)
portfolio_esg = (df_top_esg['Weight'] * df_top_esg['ESG_Total']).sum()
print(f"\nPortfolio ESG Score: {portfolio_esg:.2f}")

# ----------------------------------------------------------------
# PART C: COMPARISON WITH MARKET PORTFOLIO (SIMULATED)
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART C: ESG Portfolio vs Market Benchmark")
print("-"*60)

# Simulate market portfolio (all companies equally weighted)
df_esg['Market_Weight'] = 1 / len(df_esg)
market_esg = (df_esg['Market_Weight'] * df_esg['ESG_Total']).sum()

# Simulate financial returns (simplified: random normal)
df_esg['Expected_Return'] = np.random.normal(0.05, 0.02, len(df_esg)).clip(0.01, 0.12)
df_esg['Risk'] = np.random.normal(0.15, 0.05, len(df_esg)).clip(0.05, 0.35)

# ESG portfolio return
df_top_esg = df_top_esg.merge(df_esg[['Company', 'Expected_Return', 'Risk']], on='Company')
portfolio_return = (df_top_esg['Weight'] * df_top_esg['Expected_Return']).sum()
portfolio_risk = np.sqrt(sum((df_top_esg['Weight'] * df_top_esg['Risk'])**2))  # simplified

# Market portfolio return
market_return = df_esg['Expected_Return'].mean()
market_risk = df_esg['Risk'].mean()

print(f"ESG Portfolio: Return = {portfolio_return*100:.2f}%, Risk = {portfolio_risk*100:.2f}%")
print(f"Market Portfolio: Return = {market_return*100:.2f}%, Risk = {market_risk*100:.2f}%")
print(f"ESG Portfolio ESG Score: {portfolio_esg:.2f}")
print(f"Market Portfolio ESG Score: {market_esg:.2f}")

# ----------------------------------------------------------------
# PART D: CARBON FOOTPRINT ANALYSIS
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART D: Carbon Footprint Analysis")
print("-"*60)

# Calculate total carbon emissions for each company (tonnes)
df_esg['Total_Emissions'] = df_esg['Revenue'] * df_esg['Carbon_Intensity']

# Portfolio carbon footprint (weighted average intensity)
df_top_esg = df_top_esg.merge(df_esg[['Company', 'Carbon_Intensity', 'Total_Emissions']], on='Company')
portfolio_carbon_intensity = (df_top_esg['Weight'] * df_top_esg['Carbon_Intensity']).sum()
market_carbon_intensity = df_esg['Carbon_Intensity'].mean()

print(f"Portfolio Carbon Intensity: {portfolio_carbon_intensity:.2f} tonnes CO2 per $M revenue")
print(f"Market Carbon Intensity: {market_carbon_intensity:.2f} tonnes CO2 per $M revenue")

# Sector carbon contribution
sector_carbon = df_esg.groupby('Sector')['Total_Emissions'].sum().sort_values(ascending=False)

fig, axes = plt.subplots(2, 2, figsize=(14, 10))

# ESG Score Distribution
ax = axes[0, 0]
ax.hist(df_esg['ESG_Total'], bins=20, edgecolor='black', alpha=0.7, color='green')
ax.axvline(portfolio_esg, color='red', linestyle='--', label=f'Portfolio ESG: {portfolio_esg:.1f}')
ax.axvline(market_esg, color='blue', linestyle=':', label=f'Market ESG: {market_esg:.1f}')
ax.set_xlabel('ESG Score')
ax.set_ylabel('Frequency')
ax.set_title('ESG Score Distribution')
ax.legend()
ax.grid(True, alpha=0.3)

# Sector Allocation
ax = axes[0, 1]
sector_allocation.sort_values().plot(kind='barh', ax=ax, color='teal', alpha=0.7)
ax.set_xlabel('Portfolio Weight (%)')
ax.set_title('ESG Portfolio Sector Allocation')

# Carbon Intensity by Sector (boxplot)
ax = axes[1, 0]
sns.boxplot(data=df_esg, x='Sector', y='Carbon_Intensity', ax=ax)
ax.set_ylabel('Carbon Intensity (tonnes/$M revenue)')
ax.set_title('Carbon Intensity by Sector')
ax.tick_params(axis='x', rotation=45)

# ESG vs Financial Performance (scatter)
ax = axes[1, 1]
scatter = ax.scatter(df_esg['ESG_Total'], df_esg['Expected_Return'], 
                     c=df_esg['Market_Cap'], cmap='viridis', alpha=0.6, s=50)
ax.set_xlabel('ESG Score')
ax.set_ylabel('Expected Return')
ax.set_title('ESG vs Expected Return (size = Market Cap)')
plt.colorbar(scatter, ax=ax, label='Market Cap ($M)')
ax.grid(True, alpha=0.3)

plt.tight_layout()
plt.savefig('esg_analytics.png', dpi=300)
plt.show()

# ----------------------------------------------------------------
# PART E: REGULATORY CONTEXT
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART E: Regulatory Context")
print("-"*60)

print("""
Key Regulations:

1. SFDR (EU, 2021):
   - Financial market participants must classify products as Article 6, 8, or 9.
   - Disclosures on how sustainability risks are integrated.
   - Principal Adverse Impact (PAI) indicators.

2. CSRD (EU, 2024):
   - Expands non-financial reporting to more companies.
   - Requires double materiality assessment.
   - Aligned with ESRS (European Sustainability Reporting Standards).

3. TCFD (G20, 2017):
   - Recommendations for climate-related financial disclosures.
   - Governance, strategy, risk management, metrics & targets.

4. ISSB (IFRS, 2023):
   - IFRS S1: General requirements for sustainability disclosures.
   - IFRS S2: Climate-related disclosures.
   - Global baseline for investor-focused reporting.

5. SEC Climate Rules (US, 2024):
   - Requires disclosure of climate-related risks and greenhouse gas emissions.
   - Scope 1 and 2 for larger companies; Scope 3 for some.

6. EU Taxonomy:
   - Classification of environmentally sustainable activities.
   - Technical screening criteria for climate change mitigation and adaptation.

Implications for Banks:
  - Enhanced due diligence on borrowers' ESG risks.
  - Integration of ESG into credit and investment decisions.
  - Reporting on financed emissions (PCAF methodology).
  - Development of green products (green loans, bonds).
""")

# ----------------------------------------------------------------
# PART F: GREEN BOND AND IMPACT INVESTING
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART F: Green Bond and Impact Investing Simulation")
print("-"*60)

# Simulate a green bond portfolio
n_bonds = 20
bond_data = []
for i in range(n_bonds):
    bond_data.append({
        'Bond_ID': f'GB_{i+1}',
        'Issuer_ESG': np.random.uniform(50, 95),
        'Coupon': np.random.uniform(0.02, 0.06),
        'Maturity': np.random.choice([3, 5, 7, 10]),
        'Green_Label': np.random.choice([True, False], p=[0.7, 0.3]),
        'Project_Type': np.random.choice(['Renewable Energy', 'Energy Efficiency', 'Pollution Control', 'Sustainable Agriculture'])
    })

df_bonds = pd.DataFrame(bond_data)
green_bonds = df_bonds[df_bonds['Green_Label'] == True]
print(f"Green Bonds: {len(green_bonds)} out of {len(df_bonds)}")

# Calculate green bond premium (simplified)
green_premium = 0.005  # 50 bps lower yield for green bonds
green_bonds['Adjusted_Coupon'] = green_bonds['Coupon'] - green_premium

print("\nGreen Bond Portfolio:")
print(green_bonds.head().round(4))

# ----------------------------------------------------------------
# PART G: ESG RISK ASSESSMENT
# ----------------------------------------------------------------

print("\n" + "-"*60)
print("PART G: ESG Risk Assessment")
print("-"*60)

def esg_risk_score(df, weights={'E': 0.4, 'S': 0.3, 'G': 0.3}):
    """
    Calculate ESG risk score (higher = more risk).
    """
    # Higher E/S/G scores mean lower risk (inverse)
    e_risk = 100 - df['E_Score']
    s_risk = 100 - df['S_Score']
    g_risk = 100 - df['G_Score']
    
    total_risk = weights['E'] * e_risk + weights['S'] * s_risk + weights['G'] * g_risk
    return total_risk

df_esg['ESG_Risk_Score'] = esg_risk_score(df_esg)

# Identify high-risk companies (top 10% risk)
threshold = df_esg['ESG_Risk_Score'].quantile(0.9)
high_risk = df_esg[df_esg['ESG_Risk_Score'] >= threshold]
print(f"High-risk companies (top 10%): {len(high_risk)}")
print(high_risk[['Company', 'Sector', 'ESG_Risk_Score']].head(10).round(2))

# ----------------------------------------------------------------
# PART H: SUMMARY AND RECOMMENDATIONS
# ----------------------------------------------------------------

print("\n" + "="*70)
print("PART H: Summary and Recommendations")
print("="*70)

print("""
Sustainable Finance and ESG Analytics – Key Takeaways:

1. ESG factors are material to financial performance and risk.
2. Frameworks (GRI, SASB, TCFD, ISSB) provide guidance for disclosure and integration.
3. ESG data challenges: standardisation, quality, coverage, and greenwashing.
4. Quantitative methods: scoring, screening, integration, and impact measurement.
5. Regulatory pressure is increasing (SFDR, CSRD, SEC).
6. ESG integration can enhance risk-adjusted returns and mitigate risk.
7. Green bonds and impact investing are growing rapidly.

Recommendations for Banks and Investors:
  - Develop robust ESG data infrastructure (internal and external).
  - Integrate ESG into credit risk assessment and investment processes.
  - Use climate scenario analysis (NGFS scenarios) for stress testing.
  - Report on financed emissions using PCAF.
  - Engage with companies to improve ESG performance.
  - Offer green financial products (loans, bonds, ETFs).

Recommendations for Data Practitioners:
  - Build skills in ESG data analytics and reporting.
  - Understand regulatory requirements and disclosure standards.
  - Leverage alternative data for ESG insights.
  - Apply machine learning for ESG sentiment and controversy detection.
  - Support sustainability reporting with visualisations and storytelling.
""")

print("="*70)
print("END OF BONUS LESSON 7")
print("="*70)