Car insurance claim risk model
Inspired by publicly reported work at Progressive
Predict which drivers will file a claim next year, and check the model treats groups fairly.
The brief
Progressive has written about pricing car insurance on how people actually drive, and about which behaviours go with more claims. Start with the classic insurance version: using public policy data, predict which drivers will claim, measure the model with the Gini coefficient insurers use, and check it for unfair differences between groups. Good looks like: normalised Gini against a simple baseline, a lift chart, and a fairness check with a short written discussion.
Read the case first
What Progressive has said in public about this kind of work, on its own pages:
- Snapshot (Progressive) ↗
- Lead Foot Report: hard braking and crash risk (Progressive newsroom, 2015) ↗
- Putting the phone down correlates to lower insurance claims (Progressive newsroom, 2019) ↗
Make it your own
Set it in India: two-wheeler insurance, or Motor Vehicles Act rules on what can be used in pricing. Say which features you would not be allowed or willing to use, drop them, and report the cost.
AITeky has no link with Progressive, which has not reviewed or endorsed this brief. The brief only draws on what the company has published, uses open data rather than its data, and is a learning exercise: don't present your project as Progressive's work or system.
What you will build
- Explore the anonymised policy features and claims
- Handle missing values and imbalance
- Train logistic regression and gradient boosting
- Evaluate with normalised Gini and lift charts
- Check fairness across groups and discuss it
Free, open data to build it with
Sign up free to get the data and links
It takes a minute with your email. AiTeky Pass (₹999/month or ₹9,999/year) is free for your first 30 days.