Predict SME Business Distress

Using advanced machine learning and World Bank Enterprise Survey data to identify financial distress signals across East African SMEs

📈 Explore Tools View on GitHub
14,688
Firms Analyzed
8
Countries Covered
86.36%
ROC-AUC Score

Key Features

🎯

3-Class Prediction

Classify SMEs into Stable, Moderate Risk, or High Risk categories using 32 leakage-free features — finance access, productivity, obstacle burden, and firm maturity

🔍

SHAP Interpretability

Understand exactly which factors drive distress predictions with firm-level explanations

🌍

Cross-Country Generalisation

Single model trained on 8 countries validates genuine cross-country performance — ROC-AUC 0.86+ achieved across all 8 nations with no country-specific retraining