EnergyInfrastructureAI / seismic_analysis_python.py
dangmanhtruong's picture
Clean commit with LFS-tracked images
845d5aa
import pandas as pd
import numpy as np
import math
from typing import Dict, List, Tuple, Optional
from collections import defaultdict
import json
from os.path import join as pjoin
class SeismicDrillingAnalyzer:
"""
Analyzes seismic activity vs drilling density to identify areas with
high seismic survey density but low recent drilling activity.
"""
def __init__(self, proximity_radius_km: float = 50.0, min_earthquake_count: int = 1):
"""
Initialize the analyzer with configurable parameters.
Args:
proximity_radius_km: Radius in km to consider wells as "nearby"
min_earthquake_count: Minimum earthquakes required for a region to be analyzed
"""
self.proximity_radius = proximity_radius_km
self.min_earthquake_count = min_earthquake_count
self.uk_bounds = {
'lat_min': 49.0, 'lat_max': 61.0,
'lon_min': -8.0, 'lon_max': 3.0
}
def load_and_validate_data(self, earthquake_file: str, well_file: str) -> Tuple[pd.DataFrame, pd.DataFrame]:
"""
Load CSV files and perform data quality assessment.
Args:
earthquake_file: Path to earthquake CSV file
well_file: Path to well production CSV file
Returns:
Tuple of (earthquake_df, well_df) with validated data
"""
print("Loading data files...")
# Load earthquake data
earthquake_df = pd.read_csv(earthquake_file)
print(f"Loaded {len(earthquake_df)} earthquake records")
# Load well data
well_df = pd.read_csv(well_file)
print(f"Loaded {len(well_df)} well records")
# Clean and validate earthquake data
earthquake_df = self._validate_coordinates(earthquake_df, "earthquake")
# Clean and validate well data (with coordinate swap detection)
well_df = self._validate_coordinates(well_df, "well")
well_df = self._detect_and_fix_coordinate_swap(well_df)
return earthquake_df, well_df
def _validate_coordinates(self, df: pd.DataFrame, data_type: str) -> pd.DataFrame:
"""Validate and clean coordinate data."""
initial_count = len(df)
# Remove rows with missing coordinates
df = df.dropna(subset=['Lat', 'Lon'])
# Remove rows with invalid coordinates (non-numeric)
df = df[pd.to_numeric(df['Lat'], errors='coerce').notna()]
df = df[pd.to_numeric(df['Lon'], errors='coerce').notna()]
# Convert to numeric
df['Lat'] = pd.to_numeric(df['Lat'])
df['Lon'] = pd.to_numeric(df['Lon'])
final_count = len(df)
if final_count < initial_count:
print(f"Removed {initial_count - final_count} {data_type} records with invalid coordinates")
return df
def _detect_and_fix_coordinate_swap(self, df: pd.DataFrame) -> pd.DataFrame:
"""
Detect if Lat/Lon columns are swapped based on UK geographic bounds.
"""
lat_in_bounds = ((df['Lat'] >= self.uk_bounds['lat_min']) &
(df['Lat'] <= self.uk_bounds['lat_max'])).mean()
lon_in_bounds = ((df['Lon'] >= self.uk_bounds['lon_min']) &
(df['Lon'] <= self.uk_bounds['lon_max'])).mean()
# Check if swapping would improve coordinate validity
lat_as_lon = ((df['Lat'] >= self.uk_bounds['lon_min']) &
(df['Lat'] <= self.uk_bounds['lon_max'])).mean()
lon_as_lat = ((df['Lon'] >= self.uk_bounds['lat_min']) &
(df['Lon'] <= self.uk_bounds['lat_max'])).mean()
if (lat_as_lon > lat_in_bounds) and (lon_as_lat > lon_in_bounds):
print("Detected coordinate swap - fixing...")
df = df.copy()
df['Lat'], df['Lon'] = df['Lon'].copy(), df['Lat'].copy()
return df
def calculate_distance(self, lat1: float, lon1: float, lat2: float, lon2: float) -> float:
"""
Calculate distance between two points using Haversine formula.
Returns:
Distance in kilometers
"""
R = 6371 # Earth's radius in km
# Convert to radians
lat1, lon1, lat2, lon2 = map(math.radians, [lat1, lon1, lat2, lon2])
dlat = lat2 - lat1
dlon = lon2 - lon1
a = (math.sin(dlat/2)**2 +
math.cos(lat1) * math.cos(lat2) * math.sin(dlon/2)**2)
c = 2 * math.asin(math.sqrt(a))
return R * c
def analyze_regional_distribution(self, earthquake_df: pd.DataFrame, well_df: pd.DataFrame) -> Dict:
"""
Analyze earthquake and well distribution by region/field.
"""
print("\nAnalyzing regional distribution...")
# Count earthquakes by region
earthquake_regions = earthquake_df['Region'].value_counts().to_dict()
# Count wells by field
well_fields = well_df['Field'].value_counts().to_dict()
return {
'earthquake_regions': earthquake_regions,
'well_fields': well_fields,
'top_earthquake_regions': dict(list(earthquake_regions.items())[:10]),
'top_well_fields': dict(list(well_fields.items())[:10])
}
def analyze_spatial_proximity(self, earthquake_df: pd.DataFrame, well_df: pd.DataFrame) -> List[Dict]:
"""
Analyze spatial relationship between earthquake regions and well locations.
"""
print(f"\nAnalyzing spatial proximity (radius: {self.proximity_radius}km)...")
# Group earthquakes by region
earthquake_regions = earthquake_df.groupby('Region')
region_analysis = []
for region_name, region_group in earthquake_regions:
if len(region_group) < self.min_earthquake_count:
continue
# Calculate regional centroid
avg_lat = region_group['Lat'].mean()
avg_lon = region_group['Lon'].mean()
earthquake_count = len(region_group)
# Count nearby wells
nearby_wells = 0
well_distances = []
for _, well in well_df.iterrows():
distance = self.calculate_distance(avg_lat, avg_lon, well['Lat'], well['Lon'])
well_distances.append(distance)
if distance <= self.proximity_radius:
nearby_wells += 1
# Calculate ratio (avoid division by zero)
ratio = earthquake_count / max(nearby_wells, 1)
region_analysis.append({
'region': region_name,
'earthquake_count': earthquake_count,
'nearby_wells': nearby_wells,
'avg_lat': avg_lat,
'avg_lon': avg_lon,
'ratio': ratio,
'min_well_distance': min(well_distances) if well_distances else float('inf')
})
# Sort by ratio (high seismic activity, low drilling density)
region_analysis.sort(key=lambda x: x['ratio'], reverse=True)
return region_analysis
def create_grid_analysis(self, earthquake_df: pd.DataFrame, well_df: pd.DataFrame,
cell_size_degrees: float = 0.5) -> List[Dict]:
"""
Alternative grid-based analysis approach.
"""
print(f"\nPerforming grid-based analysis (cell size: {cell_size_degrees}°)...")
# Calculate bounds
all_lats = list(earthquake_df['Lat']) + list(well_df['Lat'])
all_lons = list(earthquake_df['Lon']) + list(well_df['Lon'])
min_lat, max_lat = min(all_lats), max(all_lats)
min_lon, max_lon = min(all_lons), max(all_lons)
# Create grid
lat_cells = int(math.ceil((max_lat - min_lat) / cell_size_degrees))
lon_cells = int(math.ceil((max_lon - min_lon) / cell_size_degrees))
grid_analysis = []
for i in range(lat_cells):
for j in range(lon_cells):
cell_min_lat = min_lat + i * cell_size_degrees
cell_max_lat = min_lat + (i + 1) * cell_size_degrees
cell_min_lon = min_lon + j * cell_size_degrees
cell_max_lon = min_lon + (j + 1) * cell_size_degrees
# Count earthquakes in cell
eq_in_cell = earthquake_df[
(earthquake_df['Lat'] >= cell_min_lat) &
(earthquake_df['Lat'] < cell_max_lat) &
(earthquake_df['Lon'] >= cell_min_lon) &
(earthquake_df['Lon'] < cell_max_lon)
]
# Count wells in cell
wells_in_cell = well_df[
(well_df['Lat'] >= cell_min_lat) &
(well_df['Lat'] < cell_max_lat) &
(well_df['Lon'] >= cell_min_lon) &
(well_df['Lon'] < cell_max_lon)
]
earthquake_count = len(eq_in_cell)
well_count = len(wells_in_cell)
if earthquake_count > 0 or well_count > 0:
ratio = earthquake_count / max(well_count, 1)
grid_analysis.append({
'grid_id': f"{i}_{j}",
'lat_range': (cell_min_lat, cell_max_lat),
'lon_range': (cell_min_lon, cell_max_lon),
'center_lat': (cell_min_lat + cell_max_lat) / 2,
'center_lon': (cell_min_lon + cell_max_lon) / 2,
'earthquake_count': earthquake_count,
'well_count': well_count,
'ratio': ratio
})
# Sort by ratio
grid_analysis.sort(key=lambda x: x['ratio'], reverse=True)
return grid_analysis
def generate_summary_report(self, regional_analysis: List[Dict],
grid_analysis: List[Dict],
distribution_stats: Dict,
top_n: int = 15) -> Dict:
"""
Generate comprehensive analysis summary.
"""
print("\nGenerating summary report...")
# Priority classification
high_priority = [r for r in regional_analysis[:top_n] if r['ratio'] >= 4.0]
medium_priority = [r for r in regional_analysis[:top_n] if 2.0 <= r['ratio'] < 4.0]
low_priority = [r for r in regional_analysis[:top_n] if r['ratio'] < 2.0]
report = {
'analysis_parameters': {
'proximity_radius_km': self.proximity_radius,
'min_earthquake_count': self.min_earthquake_count
},
'data_summary': {
'total_regions_analyzed': len(regional_analysis),
'high_priority_areas': len(high_priority),
'medium_priority_areas': len(medium_priority),
'low_priority_areas': len(low_priority)
},
'top_regions': regional_analysis[:top_n],
'priority_classification': {
'high_priority': high_priority,
'medium_priority': medium_priority,
'low_priority': low_priority
},
'top_grid_cells': grid_analysis[:10],
'distribution_stats': distribution_stats,
'key_insights': self._generate_insights(regional_analysis, distribution_stats)
}
return report
def _generate_insights(self, regional_analysis: List[Dict], distribution_stats: Dict) -> List[str]:
"""Generate key insights from the analysis."""
insights = []
# High seismic, zero drilling areas
zero_drilling = [r for r in regional_analysis if r['nearby_wells'] == 0]
if zero_drilling:
insights.append(f"{len(zero_drilling)} regions have earthquake activity but zero wells within {self.proximity_radius}km")
# Geographic patterns
irish_sea_regions = [r for r in regional_analysis if 'IRISH SEA' in r['region']]
if irish_sea_regions:
insights.append(f"Irish Sea shows {irish_sea_regions[0]['earthquake_count']} earthquakes with {irish_sea_regions[0]['nearby_wells']} nearby wells")
# Top earthquake region
if regional_analysis:
top_region = regional_analysis[0]
insights.append(f"Highest priority area: {top_region['region']} ({top_region['earthquake_count']} earthquakes, {top_region['nearby_wells']} wells)")
# Well concentration
top_field = max(distribution_stats['well_fields'].items(), key=lambda x: x[1])
insights.append(f"Most active drilling field: {top_field[0]} ({top_field[1]} wells)")
return insights
def save_results(self, report: Dict, output_file: str = "seismic_analysis_results.json"):
"""Save analysis results to JSON file."""
with open(output_file, 'w') as f:
json.dump(report, f, indent=2, default=str)
print(f"\nResults saved to: {output_file}")
def run_complete_analysis(self, earthquake_file: str, well_file: str) -> Dict:
"""
Run the complete analysis pipeline.
Args:
earthquake_file: Path to earthquake CSV file
well_file: Path to well production CSV file
Returns:
Complete analysis report dictionary
"""
print("=== UK Seismic Activity vs Drilling Analysis ===")
# Step 1: Load and validate data
earthquake_df, well_df = self.load_and_validate_data(earthquake_file, well_file)
# Step 2: Analyze regional distribution
distribution_stats = self.analyze_regional_distribution(earthquake_df, well_df)
# Step 3: Spatial proximity analysis
regional_analysis = self.analyze_spatial_proximity(earthquake_df, well_df)
# Step 4: Grid-based analysis (alternative approach)
grid_analysis = self.create_grid_analysis(earthquake_df, well_df)
# Step 5: Generate comprehensive report
report = self.generate_summary_report(regional_analysis, grid_analysis, distribution_stats)
# Step 6: Display results
final_report = self._display_results(report)
print(final_report)
return report, final_report
def _display_results(self, report: Dict):
"""Display formatted analysis results."""
output = []
output.append("\n" + "="*60)
output.append("ANALYSIS RESULTS")
output.append("="*60)
output.append(f"\nData Summary:")
output.append(f"- Total regions analyzed: {report['data_summary']['total_regions_analyzed']}")
output.append(f"- High priority areas: {report['data_summary']['high_priority_areas']}")
output.append(f"- Medium priority areas: {report['data_summary']['medium_priority_areas']}")
output.append(f"\nTop 10 Priority Areas (High Seismic Activity, Low Drilling):")
output.append("-" * 80)
output.append(f"{'Rank':<4} {'Region':<20} {'Earthquakes':<11} {'Wells':<6} {'Ratio':<6} {'Location'}")
output.append("-" * 80)
for i, region in enumerate(report['top_regions'][:10], 1):
output.append(f"{i:<4} {region['region'][:19]:<20} {region['earthquake_count']:<11} "
f"{region['nearby_wells']:<6} {region['ratio']:<6.1f} "
f"{region['avg_lat']:.3f}°N, {abs(region['avg_lon']):.3f}°W")
output.append(f"\nKey Insights:")
for insight in report['key_insights']:
output.append(f"• {insight}")
return "\n".join(output)
# Example usage and configuration
def main():
"""Example usage of the SeismicDrillingAnalyzer."""
# Initialize analyzer with custom parameters
analyzer = SeismicDrillingAnalyzer(
proximity_radius_km=50.0, # Consider wells within 50km as "nearby"
min_earthquake_count=1 # Analyze regions with at least 1 earthquake
)
base_path = "/media/dangmanhtruong/147E655C7E65379E/TRUONG/Proposal_writing/Energy_Infrastructure_AI"
well_data_path = pjoin(base_path, "datasets", "UKCS Daily Production Data", "UKCS_well_production_avg_data_processed.csv")
seismic_data_path = pjoin(base_path, "datasets", "BGS_earthquake_data", "UK_BGS_earthquate_data.csv")
# Run complete analysis
try:
report, final_report = analyzer.run_complete_analysis(
earthquake_file=seismic_data_path,
well_file=well_data_path,
)
# Save results
analyzer.save_results(report)
# Additional custom analysis examples
print("\n" + "="*60)
print("CUSTOM ANALYSIS EXAMPLES")
print("="*60)
# Example: Change proximity radius
analyzer_strict = SeismicDrillingAnalyzer(proximity_radius_km=25.0)
print(f"\nWith stricter 25km radius:")
# You could run partial analysis here
# Example: Focus on high-activity regions only
analyzer_high_activity = SeismicDrillingAnalyzer(min_earthquake_count=3)
print(f"\nFocusing on regions with 3+ earthquakes:")
# You could run partial analysis here
except FileNotFoundError as e:
print(f"Error: Could not find data file - {e}")
print("Please ensure the CSV files are in the correct location.")
except Exception as e:
print(f"Analysis error: {e}")
if __name__ == "__main__":
main()