𝜗𝜚 ࣪˖ ִ𐙚 heyyyyy divasss ✧˖°⊹ ࣪ ˖

visualize.py

Purpose: provides comprehensive data visualization and analysis for disaster-related social media posts, generating insightful charts, metrics, and interactive maps from CSV data.

Features:

• Data Preprocessing: Cleans and prepares disaster data with timestamp conversion and location handling
• Multiple Visualization Types: Creates pie charts, bar charts, line graphs, and box plots
• Interactive Geographic Mapping: Generates Folium maps with disaster location markers
• Temporal Analysis: Shows daily trends and hourly patterns of disaster reports
• Severity Assessment: Analyzes disaster severity levels with model confidence metrics
• Help Request Tracking: Identifies posts requesting assistance by disaster type

Outputs: (generates several visualization files as pngs)

• disaster_type_distribution.png - Disaster frequency by type
• severity_analysis.png - Severity levels and confidence scores
• temporal_analysis.png - Daily and hourly posting patterns
• help_request_analysis.png - Help requests across disaster types
• model_confidence_analysis.png - AI model performance by disaster type
• location_analysis.png - Top mentioned locations
• author_analysis.png - Most active reporters
• disaster_geographic_map.html - Interactive map with disaster locations

Usage:

# Initialize with your CSV file
visualizer = DisasterDataVisualizer("your_disaster_data.csv")

# Run complete analysis
visualizer.run_all_analyses()
# -- OR -->
python3 visualize.py # ensure you're in src on terminal

# Or run individual analyses
visualizer.disaster_type_distribution()
visualizer.create_geographic_map()
visualizer.generate_dashboard_metrics()

Data Requirements: The CSV file should contain columns for:

• disaster_type (categorical)
• severity_level (high/medium/low)
• location_mentioned (text)
• latitude/longitude (coordinates)
• help_request (boolean)
• model_confidence (0-1 score)
• author (text)
• processed_at (timestamp)

Dependencies:

• pandas, matplotlib, seaborn - Data analysis and static visualizations
• folium - Interactive geographic maps
• plotly - Advanced interactive charts
• numpy - Numerical computations

heatmap.py

Purpose: creates interactive geographic heatmaps from disaster data, automatically enhancing location coordinates and visualizing disaster intensity across regions.

Features:

• Automatic Geocoding: Converts location names to coordinates using OpenStreetMap's Nominatim service
• Data Enhancement: Fills in missing coordinates using location mentions
• Interactive Heatmaps: Creates layered Folium maps with both heatmap overlays and individual markers
• Intensity Calculation: Combines severity levels and model confidence for dynamic heatmap weighting
• Dual Output: Generates both comprehensive and simplified map versions

Outputs: The script generates two interactive HTML maps:

• disaster_heatmap.html - Comprehensive version with:
  • Color-coded intensity heatmap (blue → red gradient)
  • Individual disaster markers with detailed popups
  • Interactive layer controls
  • Built-in legend and instructions
  • Severity-weighted intensity calculation
• disaster_heatmap_simple.html - Streamlined version with:
  • Confidence-based heatmap
  • Visible disaster markers by default
  • Clean, focused visualization

How it Works:

1. Data Loading & Cleaning:
   • Reads CSV disaster data
   • Identifies records with missing coordinates
   • Uses geocoding to enhance incomplete data
2. Coordinate Enhancement:
   • Attempts multiple geocoding methods (geopy + direct API)
   • Adds delays to respect rate limits
   • Provides detailed progress reporting
3. Intensity Calculation:
   • Maps severity levels to weights (High: 1.0, Medium: 0.7, Low: 0.4)
   • Calculates intensity = severity weight × confidence score
   • Creates gradient from low (blue) to high (red) intensity
4. Map Generation:
   • Automatic center calculation from data
   • Color-coded markers by disaster type
   • Detailed popups with event information
   • Layer controls to toggle heatmap/markers

Usage:

# Basic usage (run the script directly)
python3 disaster_heatmap.py

# Or use the functions individually:
from disaster_heatmap import create_disaster_heatmap, print_data_summary

# Check your data first
print_data_summary()

# Generate the main heatmap
heatmap, data = create_disaster_heatmap('custom_output.html')

# Generate simplified version
simple_map = create_simple_heatmap_with_markers('simple_map.html')

Data Requirements: Same as visualize.py

Dependencies:

• folium - Interactive map generation
• geopy - Location geocoding services
• pandas - Data processing and cleaning
• requests - API calls for geocoding
• matplotlib/seaborn - Optional data analysis

Features for Emergency Response:

• Rapid Assessment: Visualize disaster concentration areas
• Severity Prioritization: Identify high-intensity zones
• Resource Allocation: See help request patterns
• Multi-layer Analysis: Toggle between overview and individual events