A collected reference for pricing teams getting started, going deeper, or evaluating the libraries against commercial alternatives.
Getting started
Getting Started — Three entry paths depending on your background: pricing actuary moving to Python, data scientist joining an insurance team, or technical team lead evaluating what to adopt. Each path recommends which libraries to reach for first and in what order.
Complete Motor Pricing Pipeline — A full production workflow in one place: synthetic data with a known DGP, GBM training, SHAP factor extraction, GLM distillation, conformal prediction intervals, fairness audit, drift monitoring, and governance pack. The right starting point if you want to see all the libraries working together.
Which Library Do I Need? — Problem-oriented decision guide. Maps a pricing challenge to the right library, organised by workflow stage from data preparation through to deployment and monitoring.
Reference
Databricks Notebooks — Ten production-ready Databricks notebooks covering the full pricing workflow. Each runs on synthetic UK motor data with no external dependencies. Download individually or as a zip.
Proxy Discrimination Audit Guide — Step-by-step compliance guide for UK personal lines teams. Covers the Consumer Duty and Equality Act 2010 obligations, how to conduct a defensible proxy audit, and what the board needs to see.
Benchmarks and comparisons
Benchmarks — Every library ships with a Databricks-runnable benchmark on a known data-generating process. Results are extracted from those runs. Failures are published alongside successes.
Open-Source vs Commercial Platforms — For pricing teams evaluating Burning Cost against Emblem, Radar, Akur8, or DataRobot. Covers what each commercial tool does, where open-source covers the same ground, and where the trade-offs lie.