Telematics-based discount: pricing model design

Ситуация: Запускаем telematics (driving data via app/device). Plan: discount safe drivers 15%, surcharge risky 10%. Как design pricing model?

Voluntary opt-in. ~10% customers willing. Data: speed, braking, mileage, time-of-day.

Доступные данные

• telematics_data: customer_id, week, speeding_events, hard_braking, mileage, night_driving_pct
• claims: customer_id, ts, amount
• policies: id, customer_id, current_premium

Задачи

• Risk score design: какие signals predictive.
• Discount/surcharge curve: где cutoffs.
• Adverse selection: какие customers opt in.
• Pricing impact: aggregate revenue effect.

Все кейсы для подготовки → ← Все кейсы