cranalytics

cranalytics is a DataFrame-first toolkit for practical credit risk work: vintage curves, lifetime loss forecasting, portfolio segmentation, feature analytics, ML modeling, and rollforward operating diagnostics.

Run cranalytics quickstart Open the Gallery

• Start Here

Best first step if you are new to the package.

cranalytics quickstart

Expect: one guided workflow, one first win, one obvious next step.

Getting Started

• Visual Tour

Best if you want to scan examples before reading tutorials.

python -m cranalytics.examples.core_vintage

Expect: concrete example commands and outcomes by workflow.

Workflow Gallery

• Pick By Problem

Best if you already know the risk question or data shape.

cranalytics rollforward-readiness your_rollforward_data.csv --output-dir rollforward_readiness_out

Expect: fast routing to the right workflow and next doc.

Choose Your Path

Core Workflows

• Vintage Curve Fitting

Fit cumulative loss by cohort.

First run: python -m cranalytics.examples.core_vintage

First win: Fitted curve and ultimate loss estimate.

Tutorial

• Lifetime Loss Forecasting

Estimate reserve from a transition matrix or loan history.

First run: python -m cranalytics.examples.core_lifetime_loss

First win: Reserve estimate on a mock portfolio.

Tutorial

• FICO Segmentation

Bucket the portfolio by score band and simple risk grade.

First run: python -m cranalytics.examples.core_segmentation

First win: FICO-band table and mix diagnostics.

Tutorial

• Feature Analytics

Rank signal before committing to model training.

First run: python -m cranalytics.examples.core_feature_analytics

First win: Ranked feature table.

Tutorial

• ML Modeling

Backtest, train, and score baseline models.

First run: python -m cranalytics.examples.core_ml_modeling

First win: Fold-level backtest summary table.

Tutorial

• Rollforward Workflow

Validate monthly aggregated rollforward data before operational use.

First run: cranalytics rollforward-readiness your_rollforward_data.csv --output-dir rollforward_readiness_out

First win: Readiness report directory and diagnostics.

Workflow Reference

Read In This Order

1. Getting Started
2. Workflow Gallery
3. Choose Your Path
4. Workflow Map

API Entry Points

• Vintage & Smoothing API
• Core API Reference
• Loss Forecasting API
• Predictive Modeling API
• Rollforward Workflow API

Optional Extras

pip install "cranalytics[ml]"
pip install "cranalytics[xgboost]"
pip install "cranalytics[lightgbm]"