📊 MASA Hackathon 2026: R-Ignite

Project: Assessing Climate-Related Risks for Financial Resilience Team Name: NextGenz Lead Researcher: Nicole Ng Zhen Tiing

---

1. Executive Summary

This project develops a quantitative climate risk framework for a reinsurance firm, focusing specifically on Malaysia and the Philippines. By utilizing historical data from the World Bank and EM-DAT, we model greenhouse gas (GHG) emission trajectories and evaluate their correlation with natural disaster insurance claims. Our research aims to provide actionable insights for long-term financial resilience, stress-testing various emission scenarios to inform capital reserve requirements and parametric insurance pricing.

---

2. Team Information

🏫 General Information

• University: University of Technology Sarawak (UTS)
• Team Name: NextGenz

👥 Team Members

Name	Role & Responsibility	Contact
Nicole Ng Zhen Tiing	Project Lead & Data Scientist	qpngjll607706@gmail.com
CHAN XIN EN	Report Formatting & Visual Optimization	xinen2811@gmail.com
VEANN FOO WEI LING	Copywriting & Data Collection	veann090909@gmail.com
AUSTIN KHO QI ZHANG	Literature Research & Risk Assessment	happy.family3266@gmail.com

---

3. Project Repository Structure

.
├── NextGenZ_report.pdf                                 # Final 10-page research report
├── temp final.py                                       # Primary Python script for modeling
├── climate_risk_dashboard_sea 8 final.html             # Interactive visualization dashboard
├── climate_risk_dashboard_sea 8final.pbix              # Power BI
├── README.md                                           # Project documentation
└── Data/
    └── WB_WDI_WIDEF.csv                                # World Bank WDI raw dataset

---

4. Technical Requirements

Environment

• Python: Version 3.9 or higher

Key Libraries

Library	Purpose
pandas	Robust data manipulation and cleaning
scikit-learn	Polynomial Regression, OLS, and SVR implementation
matplotlib	High-quality statistical visualizations
seaborn	High-quality statistical visualizations

Installation

Clone this repository and install dependencies using:

pip install pandas numpy scikit-learn matplotlib seaborn

---

5. Replication & Execution

To replicate the findings presented in our report:

1. Data Setup: Ensure WB_WDI_WIDEF.csv is placed in the root or /Data directory.

2. Run Analysis: Execute the Python script:

python climate_risk_analysis.py

3. Outputs:

   • Cross-Validation: The terminal will display the R² and MAE for the 5-fold cross-validation of the GHG model.
   • Visuals: The script generates:
     • fig1_ghg_projection.png — 2030 GHG projections
     • fig2_insurance_analysis.png — Correlation heatmap
   • Interactive Dashboard: Open climate_risk_dashboard.html in any modern web browser to simulate the 2030 stress-test scenarios.

---

6. Methodology

Stage	Description
Preprocessing	Addressed missing values in the WDI dataset via interpolation; used Centered Year values to mitigate multi-collinearity in regression models
Modeling	Applied Polynomial Regression to capture the non-linear nature of historical GHG emissions
Validation	5-fold cross-validation to ensure generalizability and prevent overfitting
Stress Testing	Simulated three scenarios — Baseline, Moderate Mitigation, and High Risk — to observe potential impacts on insurance claim volatility by 2030

---

7. Key Findings

Divergent Risk Profiles While the Philippines shows acute risk from typhoons, Malaysia exhibits a structural increase in flood-related claims linked to urbanization and land-use changes.

Financial Impact There is a statistically significant positive correlation between regional GHG intensity and the frequency of "secondary peril" claims.

Strategic Recommendation Reinsurers should pivot toward Parametric Reinsurance triggers based on specific climatic thresholds rather than traditional indemnity models.

---

8. AI Disclosure

This project utilized generative AI (Gemini 1.5 Flash) for initial code scaffolding and document structure refinement. All statistical logic, data interpretations, and final model tuning were performed and verified manually by the NextGenz team to ensure accuracy and compliance with MASA Hackathon standards.

---

© 2026 NextGenz — University of Technology Sarawak (UTS). All rights reserved.