AhmeddAladdin/Used-cars-prediction-by-ML

Machine Learning project to predict Used cars in Egypt # Used Cars Price Prediction (Egypt) **Predicting used car prices using machine learning** --- ## Project summary This repository contains a complete machine-learning pipeline to predict used-car prices (targeted for the Egyptian market). The project includes data cleaning and exploration, feature engineering, model training & evaluation, and a small inference script to get predictions from a saved model. The goal is to produce a reliable regression model that estimates a fair price for a used car given its attributes (make, model, year, mileage, engine/horsepower, transmission, location, etc.), plus provide reproducible code and instructions. --- ## What I used - **Language & environment:** Python 3.8+ - **Main libraries:** pandas, numpy, scikit-learn, xgboost, joblib, matplotlib, seaborn, jupyter - **Development tools:** Jupyter Notebooks for EDA and experiments; `src/` scripts for reusable code; `req.txt` for dependencies - **Optional frontend:** simple demo app (Flask or lightweight UI) — located in `frontend/` if present --- ## What I did (high level) 1. **Data collection & loading** - Collected used-car listings dataset(s) (CSV files) with fields such as `make`, `model`, `year`, `mileage`, `engine_capacity`, `fuel_type`, `transmission`, `location`, `price`, `currency`, etc. - Stored raw files under `data/raw/`. 2. **Exploratory Data Analysis (EDA)** - Studied distributions, outliers, and missing-data patterns. - Visualised relationships between features and price (year vs price, mileage vs price, brand effects). 3. **Data cleaning** - Normalised text fields (lowercasing, trimming). - Converted currencies / unified price units if needed. - Handled missing values: domain-informed imputation for numeric fields, `unknown` for categorical where appropriate. - Removed obvious outliers and duplicates. 4. **Feature engineering** - Extracted age of car: `age = current_year - year`. - Created mileage-per-year and bin categorical variables for price brackets. - …