🌱 Lifestyle ESG Credit Scoring Model for Small Businesses

🌱 Credit Evaluation Alternative Model for Low-credit Small Business Owners, Utilizing Lifestyle-based ESG Quantitative Indicators

Team Project for 2025 FIN:NECT Challenge | Jun 2025 – Aug 2025

Developed an alternative credit scoring model incorporating everyday ESG practices (e.g., electricity usage, local product sales, electronic receipts, waste reduction) to address financial blind spots of micro-enterprises.
Designed a multi-dimensional scoring framework (E: 30%, S: 25%, G: 25%, Special ESG: 20%) balancing fairness and sector-specific characteristics.
Built data pipelines with Python, Pandas, and scikit-learn, integrating public datasets, MyData APIs, and business records.
Applied ensemble learning (XGBoost, CatBoost) to predict creditworthiness, and used SHAP (SHapley Additive exPlanations) to ensure model interpretability and fairness.
Proposed policy adoption strategies for regional banks, linking ESG practice data with financial services to expand financial accessibility.
Selected as a National Finalist & Awardee in the 2025 FIN:NECT Challenge, recognized for technical innovation and policy applicability.