Credit Risk Python Packages

autocarver

Automatic optimal discretization pipeline

toad

ESC Team's credit scorecard tools.

supermodelingfactory

Production-grade Python toolkit for credit risk modeling — scorecards, LR, LightGBM/XGBoost, WOE/IV, PSI, KS, SHAP. Source-protected wheels via Cython.

compileml

Compile tree-ensemble models into deterministic, auditable decision artifacts — score, calibrated PD, risk bands, and adverse-action reason codes in one hashed JSON that produces identical integers in Python, SQL, and COBOL.

creforge

PII-safe synthetic credit bureau data from explicit behavioural rules

treecf

Constrained, threshold-aware counterfactual explanations for tree ensembles (XGBoost, LightGBM, CatBoost, sklearn) on a bundled Rust engine

fintfm

A tabular foundation model for corporate credit risk, pretrained only on synthetic data and measured in public, including where it loses.

scorecardpipeline

scorecardpipeline封装常用的风控策略分析和评分卡建模相关组件，支持pipeline式端到端评分卡建模、三方数据分析、规则集效果评估、特征有效性分析、excel报告输出、评分卡PMML导出、全流程超参数搜索等功能。核心功能：评分卡，策略分析，风控，规则挖掘，特征筛选，自动分箱

probcal

Universal post-hoc probability calibration for binary classifiers: methods, metrics, diagnostics, and auditable offsetting — numpy-only.

skorecard

scikit-learn compatible tools for building credit risk acceptance models

mobpy

Monotonic Optimal Binning algorithm is a statistical approach to transform continuous variables into optimal and monotonic categorical variables.

pamir

Open benchmark for credit-default prediction with scarce and delayed labels: 19 public datasets, label-delayed streaming and i.i.d. evaluation, leakage-controlled synthetic augmentation. pip install pamir

covenants

pytest for credit-model governance: test a credit model's documented claims (reason codes, monotonicity, features, exclusions) in CI.

tanml

Automated validation toolkit for tabular ML models in finance and regulated domains.

financialkit

Quantitative-finance toolkit for learning and teaching: options, fixed income, rates, credit, Monte Carlo, PDE, stochastic volatility, risk, portfolios -- built on numpy/scipy.

closed-loop-default-detection

Deterministic selective-labels harness: measures PD models on the declined population real data never sees, and prices it in profit. Plants ground truth; the operating frontier is a 25-seed distribution, not a point. sklearn-only, byte-reproducible.

psi-inference

Critical values and p-values for population stability testing: Yurdakul-Naranjo PSI, the Population Resemblance Statistic, and the effect-size and overlapping tests. Pure numpy and scipy.

roll-rate-analysis

Roll Rate Analysis python package. Both month over month and snapshot roll rate functionalities are supported. It utilizes Polars library for optimization and speed.

creditsetu

Open-source, explainable credit intelligence engine for thin-file and alternative-data lending. pip install creditsetu — score customers with no bureau history from behavioral transaction data, explain every decision, adapt currency/policy via config.

transitionmatrix

Statistical analysis and visualization of state transition phenomena

credit-risk

A package for credit risk analytics

driftkit

Correct, dependency-light population stability metrics for tabular models — PSI, CSI, WOE/IV and calibration drift. numpy + pandas only.

derivus

An XVA, pricing library built from the ground up with AAD in mind implemented in PyTorch

rejectkit

Reject inference for credit scoring — 8 scikit-learn-compatible methods plus an honest benchmark that tells you whether it actually helps.