example_analysis.py

#!/usr/bin/env python3
"""
Example script demonstrating the improved Himalayan expedition analysis.

This script shows how to use the new modular analysis functions with
cross-validation as recommended in the project README.

Usage:
    python example_analysis.py
"""

import pandas as pd
import numpy as np
from analysis_utils import load_and_analyze_expeditions, ExpeditionAnalyzer


def main():
    """Run the complete analysis pipeline with improvements."""
    print("=" * 60)
    print("Himalayan Expedition Success Analysis - Improved Version")
    print("=" * 60)
    
    try:
        # Run the complete analysis
        data, results, analyzer = load_and_analyze_expeditions()
        
        print("\n" + "=" * 50)
        print("SUMMARY OF IMPROVEMENTS")
        print("=" * 50)
        
        print("\n1. CROSS-VALIDATION RESULTS (addressing README recommendation):")
        print(f"   • Mean CV Accuracy: {results['cv_accuracy_mean']:.3f} "
              f"(±{results['cv_accuracy_std']*2:.3f})")
        print(f"   • Mean CV ROC AUC: {results['cv_roc_auc_mean']:.3f} "
              f"(±{results['cv_roc_auc_std']*2:.3f})")
        print("   • This provides more robust performance estimates than single train-test split")
        
        print("\n2. DATA VALIDATION:")
        print("   • Automatic validation of data structure and quality")
        print("   • Warnings for high missing data percentages")
        print("   • Error checking for invalid values")
        
        print("\n3. SPECIALIZED EVALUATION METRICS:")
        print(f"   • Peirce Skill Score (PSS): {results['pss']:.3f}")
        print(f"   • Heidke Skill Score (HSS): {results['hss']:.3f}")
        print("   • These are domain-appropriate metrics for binary classification")
        
        print("\n4. MODULAR CODE ORGANIZATION:")
        print("   • Reusable ExpeditionAnalyzer class")
        print("   • Separated data processing, validation, and modeling functions")
        print("   • Consistent plotting and output management")
        
        print("\n5. REPRODUCIBILITY ENHANCEMENTS:")
        print("   • requirements.txt file for dependency management")
        print("   • Consistent random seeds throughout analysis")
        print("   • Automated testing framework")
        
        print("\nCompare these results with the original notebook analysis:")
        print(f"   Original reported accuracy: ~78%")
        print(f"   Improved CV accuracy: {results['cv_accuracy_mean']:.1%}")
        print(f"   Original AUC: ~0.79")
        print(f"   Improved CV AUC: {results['cv_roc_auc_mean']:.3f}")
        
        # Feature importance summary
        print("\n" + "=" * 50)
        print("KEY FINDINGS CONFIRMED")
        print("=" * 50)
        
        # Get top features (would need feature ranking from analysis)
        X = data.drop(columns=['success'])
        y = data['success']
        feature_ranking = analyzer.rank_features(X, y)
        
        print("\nTop 5 most important features:")
        for i, (_, row) in enumerate(feature_ranking.head().iterrows()):
            print(f"   {i+1}. {row['Feature']}: F-score = {row['F-Score']:.2f}")
        
        print("\nThese findings support the original conclusions about:")
        print("   • Importance of supplemental oxygen use")
        print("   • Effect of route standardization")
        print("   • Seasonal climbing patterns")
        
    except FileNotFoundError:
        print("ERROR: data/expeditions.csv not found!")
        print("\nTo run this analysis, ensure you have:")
        print("1. The data/expeditions.csv file in the project directory")
        print("2. Installed the required dependencies: pip install -r requirements.txt")
        return 1
    
    except Exception as e:
        print(f"ERROR during analysis: {e}")
        return 1
    
    print("\n" + "=" * 60)
    print("Analysis complete! Check the overleaf/ directory for visualizations.")
    print("=" * 60)
    
    return 0


def demonstrate_validation():
    """Demonstrate the data validation features."""
    print("\n" + "=" * 40)
    print("DEMONSTRATING DATA VALIDATION")
    print("=" * 40)
    
    analyzer = ExpeditionAnalyzer()
    
    # Create sample problematic data
    problematic_data = pd.DataFrame({
        'mbrs_summited': [1, 2, None, 0],  # Missing value
        'max_elev_reached': [8848, -100, 7000, 6000],  # Negative value
        'season': ['Spring', 'Winter', 'Autumn', 'Summer'],
        'bad_column': [1, 2, 3, 4]  # Will be filtered out
    })
    
    print("\nValidating problematic dataset:")
    validation = analyzer.validate_data(problematic_data)
    
    print("\nValidation Results:")
    if validation['errors']:
        print("ERRORS:")
        for error in validation['errors']:
            print(f"  ❌ {error}")
    
    if validation['warnings']:
        print("WARNINGS:")
        for warning in validation['warnings']:
            print(f"  ⚠️  {warning}")
    
    if validation['info']:
        print("INFO:")
        for info in validation['info']:
            print(f"  ℹ️  {info}")


if __name__ == "__main__":
    exit_code = main()
    
    # Demonstrate additional features
    demonstrate_validation()
    
    exit(exit_code)