Car Price Prediction is a machine learning project that predicts used car prices based on vehicle attributes such as year, kilometers driven, fuel type, transmission, owner type, and other structured listing details.
The project applies data cleaning, feature engineering, exploratory data analysis, preprocessing pipelines, model training, cross-validation, and regression model evaluation to estimate used car prices.
Used car prices depend on several factors such as vehicle age, mileage, fuel type, transmission, owner history, and market demand. Manual price estimation can be inconsistent and subjective.
This project uses machine learning to build a data-driven car price prediction model. The goal is to estimate used car prices more accurately using historical listing data and vehicle specifications.
Estimating the selling price of a used car is challenging because different technical and market-related factors influence the final price.
The objective of this project is to develop a machine learning regression model that predicts used car prices based on structured car listing data.
- Analyze used car listing data
- Clean and preprocess raw car data
- Handle missing values and inconsistent entries
- Remove outliers using the IQR method
- Engineer useful features such as car age, log-transformed kilometers, and high-mileage flag
- Perform exploratory data analysis
- Train and compare multiple regression models
- Evaluate models using RMSE, MAE, and R² score
- Select the best-performing model based on evaluation results
The dataset used in this project is:
honda_car_selling.csv
The dataset contains used car listing details.
| Feature | Description |
|---|---|
| Year | Manufacturing year of the car |
| Kms Driven | Total kilometers driven |
| Fuel Type | Type of fuel used by the car |
| Transmission | Manual or automatic transmission |
| Owner Type | Ownership history of the car |
| Price | Target variable representing selling price |
Several useful features were created from the original dataset:
| Engineered Feature | Description |
|---|---|
| Age | Calculated as 2025 - Year |
| log_kms | Log-transformed kilometers driven using log(1 + Kms Driven) |
| high_mileage | Binary flag indicating whether a car has high mileage |
| Encoded categorical features | Fuel type, transmission, owner type, and other categorical variables converted using One-Hot Encoding |
The following preprocessing steps were performed:
- Removed unwanted characters from numeric columns
- Converted price and kilometers driven into numeric format
- Handled missing values using median and mode imputation
- Removed duplicate and inconsistent rows
- Removed outliers using the IQR method
- Standardized column names
- Applied One-Hot Encoding for categorical variables
- Scaled numerical features using StandardScaler
- Built preprocessing pipelines to avoid data leakage
The following regression models were trained and compared:
- Linear Regression
- Ridge Regression
- Lasso Regression
- Random Forest Regressor
- HistGradientBoosting Regressor
The models were evaluated using:
| Metric | Description |
|---|---|
| MAE | Mean Absolute Error |
| MSE | Mean Squared Error |
| RMSE | Root Mean Square Error |
| R² Score | Measures how well the model explains price variation |
The final selected model is:
Linear Regression
Linear Regression performed best among the tested models based on evaluation metrics such as RMSE, MAE, and R² score. The dataset showed strong linear relationships between important features such as car age, kilometers driven, and price, making Linear Regression suitable for this project.
- Python
- Jupyter Notebook
- Pandas
- NumPy
- Matplotlib
- Seaborn
- Scikit-learn
Car-Price-Prediction/
│
├── README.md
├── requirements.txt
├── .gitignore
│
├── data/
│ ├── README.md
│ └── honda_car_selling.csv
│
├── notebooks/
│ ├── README.md
│ └── car_price_prediction.ipynb
│
├── reports/
│ ├── README.md
│ ├── car_price_prediction_report.pdf
│ └── car_price_prediction_presentation.ppt
│
└── images/
├── README.md
├── price_distribution.png
├── log_price_distribution.png
├── correlation_heatmap.png
├── kms_driven_vs_price.png
└── age_vs_price.png
This graph shows the distribution of original used car prices in the dataset.
This graph shows the distribution of car prices after applying log transformation. Log transformation helps reduce skewness and makes the target variable more suitable for regression models.
This heatmap shows the correlation between numerical features such as kilometers driven, price, and age. It helps identify relationships between variables.
This scatter plot shows the relationship between kilometers driven and car price. It helps analyze how mileage affects used car resale value.
This scatter plot shows the relationship between car age and selling price. Older cars generally tend to have lower prices.
git clone https://github.com/Althafk7171/Car-Price-Prediction.git
cd Car-Price-Predictionpip install -r requirements.txtjupyter notebookOpen:
notebooks/car_price_prediction.ipynb
Run all cells to reproduce the analysis and model results.
The project compared Linear Regression, Ridge Regression, Lasso Regression, Random Forest Regressor, and HistGradientBoosting Regressor.
Linear Regression achieved the best performance among the tested models based on evaluation metrics such as RMSE, MAE, and R² score.
Key factors influencing used car prices include:
- Vehicle age
- Kilometers driven
- Fuel type
- Transmission type
- Ownership history
This repository includes:
- Final project report
- Project presentation
- Jupyter Notebook
- Dataset
- Visualization images
- The model is trained on historical used car listing data.
- Real-time market demand, offers, and location-based pricing are not included.
- The dataset may not include all brands and car models.
- The model performance depends on the quality and size of the dataset.
- The project currently focuses on notebook-based prediction and does not include a deployed web application.
- Deploy the model using Streamlit or Flask
- Add real-time used car market data
- Include more car brands and models
- Add features such as engine power, service history, accident records, and location
- Try advanced boosting models such as XGBoost and CatBoost
- Add hyperparameter tuning
- Build a user-friendly car price prediction web app
- Add feature importance visualization
- Muhammed Althaf K
This project is developed for academic and learning purposes.