Credit risk topic
Build reliable point-in-time datasets, labels, joins, and controls.
Build point-in-time-safe credit-risk features from payment behavior, balances, delinquencies, categories, trends, and missing data. Includes a worked synthetic example, browser-safe Python, leakage controls, and a production monitoring checklist.
Create traceable obligor, facility, application, and collateral relationships before modeling.
Separate feature availability from future outcomes to prevent look-ahead leakage.
Treat data quality as a measurable property of the model population, not a one-time cleanup.
Validate on later periods and keep resampling or feature selection inside the training process.