Credit Risk Dataset — 50k Loans, 10 Sectors: Archival Copy Used in a Bachelor Thesis

This record contains an archival copy of the dataset “Credit Risk Dataset — 50k loans, 10 sectors”, originally created by Sergey Nefedov and published on Kaggle under the MIT License. The dataset was used as the data source for the practical part of the bachelor thesis “Návrh a vytvoření prediktivního modelu pomocí strojového učení a jeho využití v AI asistentovi pro podporu rozhodování ve firmě” by Jakub Dvořák at the Faculty of Informatics and Statistics, Prague University of Economics and Business. The archived files were downloaded from the original Kaggle record and have not been modified. The purpose of this Zenodo record is to provide a stable and publicly accessible copy of the dataset used in the bachelor thesis. Original dataset:Credit Risk Dataset — 50k loans, 10 sectors Original creator:Sergey Nefedov Original platform:Kaggle Original dataset identifier:sergionefedov/credit-risk-dataset-50k-loans-10-sectors Original license:MIT License The archive contains the following files: loan_portfolio.csv – 50,000 records and 24 attributes; the main file used for modelling credit_ratings.csv – annual credit-rating migrations credit_risk_column_descriptions.txt – supplementary file containing descriptions of dataset columns macro_stress_scenarios.csv – macroeconomic stress scenarios portfolio_metrics.csv – monthly portfolio risk metrics vintage_analysis.csv – cumulative default curves by origination vintage