Multi-domain Contextual Risk Index

Description

This repository contains the implementation of a comprehensive multilayer assessment framework for generating contextual risk indices in urban environments. The system integrates multiple geospatial data layers including risk, spatiotemporal vulnerability, response capacity, and socioeconomic indicators to produce a unified contextual risk assessment. Experiments were conducted for Lisbon and Porto, Portugal.

The framework employs hexagonal grid-based spatial analysis with 100-meter edge length cells, combining:

• Response Layer (L1-Response): Mitigation zones and emergency response capacity
• Risk Layer (L2-Flood): Flood zones and environmental hazards
• Vulnerability Layer (L3-Vulnerability): Spatiotemporal vulnerability assessment across 4 scenarios
• Socioeconomic Layers (L4-L5): Gini index and median income distribution

Article Information

[TBD]

Repository Structure

/md-contextual-risk-index
├── data/                          # Geospatial datasets and processed layers
│   ├── cities/                    # City boundary files (Lisbon, Porto)
│   ├── figures/                   # Generated visualizations and plots
│   ├── multi_layers/              # Integrated multi-layer datasets
│   ├── poti/                      # POTI (Points of Interest) datasets
│   ├── response_layer/            # Mitigation zones data
│   ├── risk_layer/                # Flood zones and environmental risk data
│   ├── socioeconomic_layer/       # Gini index and income data
│   ├── time_windows/              # Temporal analysis configurations
│   ├── vulnerability_layer/       # Spatiotemporal vulnerability 
├── notebook/                      # Jupyter notebooks for analysis pipeline
├── requirements.txt               # Python dependencies
├── .gitignore                     # Git ignore patterns
├── LICENSE                        # Repository license
└── README.md                      # This documentation

Installation

Prerequisites

• Python 3.8 or higher
• Jupyter Notebook/Lab
• Git

Setup Instructions

1. Clone the repository:

   git clone git@github.com:les2feup/md-contextual-risk-index.git
   cd md-contextual-risk-index

2. Create and activate virtual environment:

   python -m venv venv
   source venv/bin/activate  # On macOS/Linux
   # or
   venv\Scripts\activate     # On Windows

3. Install dependencies:

   pip install -r requirements.txt

4. Launch Jupyter Notebook:

   jupyter notebook

Reproducing Results

Follow these notebooks in sequence to reproduce the complete analysis pipeline:

1. Spatiotemporal Vulnerability Assessment

jupyter notebook notebook/01-spatiotemporal_vulnerability_assessment.ipynb

Generates vulnerability layers for 4 temporal scenarios using POTI data and time windows.

2. Data Layer Preparation

# Convert city zones data
jupyter notebook notebook/02-convert_from_cityzones.ipynb

# Process air quality indices (Lisbon)
jupyter notebook notebook/03-lisbon_air_quality_index.ipynb

# Generate socioeconomic layers (Porto)
jupyter notebook notebook/04_gini_index_porto.ipynb
jupyter notebook notebook/05_median_income_porto.ipynb

3. Integrated Grid Generation

jupyter notebook notebook/06-integrated_hexagonal_grid_porto.ipynb

Creates grid maps integrating all data layers for spatial analysis.

4. Multi-layer Assessment and CRI Calculation

# Multi-layer integration
jupyter notebook notebook/07_multi_layer_assessment.ipynb

# Contextual Risk Index analysis
jupyter notebook notebook/08_CRI_Analysis_All_Scenarios.ipynb

Expected Outputs

• Integrated grid maps with all layers
• Contextual Risk Index calculations for all vulnerability scenarios
• Comparative visualizations and statistical analyses
• Geospatial datasets in GeoJSON format

Usage

Single Scenario Example - Load Integrated Dataset

import geopandas as gpd
import matplotlib.pyplot as plt

# Load integrated multi-layer dataset
lisbon_data = gpd.read_file('data/multi_layers/Lisbon_multi_layer_all_scenarios_with_CRI.geojson')

# Plot CRI for scenario 1
fig, ax = plt.subplots(figsize=(12, 10))
lisbon_data.plot(column='CRI_s1', cmap='RdYlGn_r', ax=ax, legend=True)
plt.title('Contextual Risk Index - Scenario 1 (Lisbon)')
plt.show()

Multi-scenario Analysis

# Compare CRI across all scenarios
scenarios = ['s1', 's2', 's3', 's4']
cri_columns = [f'CRI_{s}' for s in scenarios]

# Calculate correlation matrix
correlation_matrix = lisbon_data[cri_columns].corr()
print("CRI Correlation Matrix Across Scenarios:")
print(correlation_matrix)

Key Features

• Hexagonal Grid Analysis: 100m edge length hexagonal spatial discretization for more localized assessment
• Multi-layer Integration: Seamless combination of multiple data layers
• Temporal Scenarios: allow different spatiotemporal scenarios
• Multi City Analysis: Examples of analysis for Lisbon and Porto, Portugal
• Reproducible Pipeline: End-to-end Jupyter notebook workflow avaliable
• Geospatial Outputs: Standards-compliant GeoJSON format exports

Dependencies

Key Python packages:

• geopandas: Geospatial data processing
• matplotlib: Data visualization
• pandas: Data manipulation
• shapely: Geometric operations
• contextily: Basemap integration
• seaborn: Statistical visualizations
• verus: Hexagonal grid generation
• osmnx: OpenStreetMap data access

See requirements.txt for complete dependency list.

License

This project is licensed under the MIT License - see the LICENSE file for details.

Acknowledgments

• LES2 FEUP Research Group
• Portuguese cities of Lisbon and Porto for data availability
• OpenStreetMap contributors for geospatial data

---

Research Group: Laboratory of Software Engineering and Engineering (LES2)
Institution: Faculty of Engineering, University of Porto (FEUP)
Contact: [Contact information to be added]