bogbeideb/nigeria-informal-credit-scoring

# Credit Scoring Model for the Nigerian Informal Sector This repository contains the Python code used in the MSc Data Science dissertation: "Credit Scoring Model for the Nigerian Informal Sector Using Mobile and Transaction Data" ## Project Overview The project develops a machine-learning-based credit scoring model using mobile phone usage and transaction data to assess creditworthiness among informal-sector participants in Nigeria. ## Methods - Data preprocessing and cleaning - Feature engineering (ARI, ISS, DAR, FIR, BBI) - Logistic Regression, Random Forest, and XGBoost models - Model evaluation using ROC-AUC, Precision, Recall, and F1-score - Fairness and bias assessment across gender and region ## Data All datasets used are either publicly available or synthetically generated. No personally identifiable information (PII) is included. ## Reproducibility Python version: 3.10+ Install dependencies: pip install -r requirements.txt