### Goal

Build a regression model that predicts the **asking price** (in EUR) of a used car listed on a popular Cyprus classifieds website, based on features like year, mileage, engine size, fuel type, body type, power, location, and extras.

### Context

The dataset contains **~730 real car listings** from Cyprus. Car prices span a huge range — from under 1,000 to over 200,000 EUR — so your model will be evaluated on **Mean Absolute Percentage Error (MAPE)**, which treats a 50% error on a cheap car and a 50% error on an expensive car equally.

The data is messy: some columns have missing values, mixed formats (e.g. engine size as `"1,6L"`), and free-text fields that need parsing. Feature engineering will be key to a strong solution.

### Evaluation

Submissions are scored on **MAPE** (lower is better). See the Evaluation tab for the full scoring formula.

### Acknowledgements

Data sourced from publicly available car listings on [Bazaraki](https://www.bazaraki.com/), a popular classifieds website in Cyprus. This is a learning competition — no monetary prizes are awarded.