Prompt
Turkish Car Valuation Platform Design
Use this when you need to design a data-driven car valuation platform for a volatile market like Turkey, incorporating robust statistics and explainability.
How to use it
- Copy the prompt and paste it into ChatGPT, Claude, Gemini or any other AI.
- Replace every {{placeholder}} with your own details, or let the AI ask you for them.
- Use the follow-ups below to go deeper.
Role You are a senior product engineer and data scientist team. Your purpose is to design a reliable, transparent car valuation platform for the Turkish market, addressing high volatility, manipulated listings, and regional dynamics.
Context you provide
- {{market_dynamics}} (optional): Specific information about Turkish inflation, taxes, exchange rates, or typical price manipulation patterns.
- {{target_users}}: Individual owners, buyers, sellers.
Instructions
- Design a valuation pipeline with these stages: data ingestion (multiple noisy sources), data cleaning (outlier removal, mileage normalization), feature weighting (mileage decay, age depreciation, damage penalty, city adjustment), robust price estimation (median/trimmed mean, output range with confidence), and explainability layer.
- Support mandatory inputs: brand, model, year, fuel type, transmission, mileage, city, damage status, ownership count.
- Default tech stack: React/Next.js frontend, Python FastAPI backend, Pandas/NumPy for data, Scikit-learn for lightweight ML (rule-based + statistical hybrid).
- Justify key decisions (e.g., why median over mean, how to handle noise).
- Output a price range: lower bound (quick sale), fair market value, upper bound (optimistic), with a confidence score and explanation of what increased/decreased value.
Output format Architecture design document with pipeline steps, tech stack choices, input/output specifications, and a sample output for a given car. Tone: technical, clear, justified.
Guardrails
- Do not blindly trust listing prices; use statistical filtering.
- Avoid black-box models initially—prefer interpretable methods.
- Account for regional price differences and market volatility.
Example Input: 2018 Renault Megane, 50,000 km, Istanbul, petrol, automatic, minor damage. Output: Lower 190k TL, Fair 210k TL, Upper 225k TL, Confidence 75%. Explanation: age and mileage aligned, city premium +5%, damage penalty -8%.