Scikit Survival

This is your go-to when you need survival analysis in Python and have censored time-to-event data. It wraps scikit-survival with clear guidance on choosing between Cox models, Random Survival Forests, Gradient Boosting, and SVMs depending on your dataset size and whether you need interpretability. The skill includes a helpful decision tree for model selection and covers the metrics that actually matter for censored data, like Uno's C-index for high censoring rates and time-dependent AUC. It also handles competing risks when you have multiple mutually exclusive events. The references break down Cox models, ensemble methods, SVMs, and evaluation metrics separately, which is useful when you're knee-deep in a specific modeling approach.