credit-default

Machine Learning project for credit default prediction using Ensemble Learning models and performance comparison.

Investigations into the credit default dataset

End-to-end credit default risk pipeline on Databricks: Bronze/Silver/Gold via Spark Declarative Pipelines, MLflow-tracked XGBoost + LR models, SHAP interpretability, benchmarked against Yeh & Lien (2009).

End-to-end Credit Risk Analysis using SQL, Python, PostgreSQL, Logistic Regression, Decision Tree, Random Forest and XGBoost.

Prédiction du risque de non-remboursement d'un prêt pour la société Home Credit Group.

This project is about credit payment defaults prediction using machine learning techniques based solely on payment data. By analyzing historical payment patterns, we aim to identify customers at risk of defaulting on their loans or credit payments, helping financial institutions in risk mitigation

Credit Default Prediction

Previsione dell'insolvenza creditizia sul dataset UCI Credit Card Default: EDA, regressione logistica, k-NN e Random Forest. Progetto per il corso di Introduzione al Pensiero Computazionale e alla Data Science (UniBo, a.a. 2025/26).

An end-to-end analysis of credit card default risk. I used Python for data cleaning and EDA, and built a two-page interactive dashboard in Tableau to present my findings and business recommendations. The live dashboard is linked in the README.

Credit card default prediction model in Python's Scikit-Learn --- A comparison of logistic regression, random forest and XGBoost models

End-to-end credit default prediction using Logistic Regression, LightGBM, SHAP explainability, calibration and risk segmentation on 307K+ applicants.