TITLE “User Rental Frequency & Pricing Sensitivity Prediction + Growth Decision Simulation” SECTION 1 — DATA SETUP 1. rentals.csv rental_id | account_id | creation_time | total_amount | distance_km | rental_minutes | is_campaign | city 2. users.csv account_id | register_date | customer_segment | age | gender | city | is_getir_migrated 3. pricing.csv city | base_price | km_price | minute_price | surge_multiplier | date 4. events.csv (telematics) event_id | rental_id | event_type | event_timestamp | lat | lon SECTION 2 — BUSINESS PROBLEM One of RentalX strategic goals for 2025 is: “Increase rental frequency and offer differentiated campaigns based on users’ price sensitivity.” You are expected to: 1. Segment all users into 3 groups: • High Frequency Drivers • Mid Frequency • Low Frequency / Churn Risk 2. Predict price sensitivity score for each user (e.g., behavioral change after price increases). 3. Combine the 2 scores to answer strategic questions such as: • “Which users should receive a campaign?” • “Will this campaign increase revenue + LTV?” 4. Provide an analytical report + model demo answering the questions below. SECTION 3 — QUESTIONS A) Data Understanding & Feature Engineering 1. Which features would you engineer for price sensitivity? Examples: • impact during price increase periods • campaign usage behavior • distance/price ratio • rental elasticity B) Predictive Modeling You are expected to build two models: Model 1 — Rental Frequency Prediction (Regression / Classification) Goal: Predict the number of rentals the user will make in the next 30 days. Explain: • Which model you choose • Why this model is suitable • How model performance should be measured Model 2 — Price Sensitivity Scoring (Elasticity Model) Goal: Measure how sensitive a user is to price changes. Possible methods: • Log-log regression • Causal Impact • Difference-in-Differences • Bayesian Structural Time Series The candidate must justify the chosen method. C) Simulation & Business Decisioning Question: If we apply a 10% price decrease: • In which user segments will rental count increase? • What will be the total impact on Revenue + LTV? • Should the discount be applied in all cities or only in selected segments?