CodeAlpha_CarPricePrediction

Machine Learning model to predict car selling prices using Random Forest Regressor with feature engineering and data visualization.

Car Price Prediction with Machine Learning 🚗

Project Overview

Machine Learning model to predict used car selling prices using Random Forest Regression with feature engineering, exploratory data analysis, visualization, and model comparison.

Dataset Information

• Total Records: 303
• Source: Kaggle — Car Price Dataset

Features in Dataset

• Car_Name
• Year
• Selling_Price (Target Variable)
• Present_Price
• Driven_kms
• Fuel_Type
• Selling_type
• Transmission
• Owner

Technologies Used

• Python
• Pandas
• NumPy
• Matplotlib
• Seaborn
• Scikit-Learn

Project Workflow

1. Data Loading
2. Data Exploration
3. Data Visualization
4. Feature Engineering
5. Data Preprocessing
6. Train-Test Split
7. Model Training
8. Model Evaluation
9. Feature Importance Analysis
10. New Car Price Prediction

Visualizations Created

• Selling Price Distribution
• Cars by Fuel Type
• Selling Price by Fuel Type
• Correlation Heatmap
• Year vs Selling Price Scatter Plot
• Actual vs Predicted Price Plot
• Feature Importance Bar Chart

Feature Engineering

A new feature was created:

• Car_Age = Current Year - Manufacturing Year

This feature improved the model's ability to understand vehicle depreciation over time.

Machine Learning Models Used

Linear Regression

Used as a baseline regression model.

Random Forest Regressor

Selected as the final model due to superior predictive performance.

Parameters:

• n_estimators = 100
• random_state = 42

Model Comparison

Metric	Linear Regression	Random Forest
R² Score	84.67%	95.99%
RMSE	1.88 Lakhs	0.96 Lakhs
MAE	1.22 Lakhs	0.64 Lakhs

Random Forest significantly outperformed Linear Regression and was selected as the final model.

Results

Metric	Result
Final Model	Random Forest Regressor
R² Score	95.99%
RMSE	0.96 Lakhs
MAE	0.64 Lakhs

Key Insights

• Present Price is the most influential feature in predicting car value.
• Diesel vehicles generally have higher resale values than Petrol vehicles.
• Car age significantly affects resale price.
• Vehicles with lower mileage tend to have higher market value.
• First-owner vehicles generally command better resale prices.
• Random Forest delivered highly accurate predictions with strong generalization performance.

Skills Demonstrated

• Data Cleaning
• Exploratory Data Analysis (EDA)
• Data Visualization
• Feature Engineering
• Regression Modeling
• Model Evaluation
• Feature Importance Analysis
• Predictive Analytics
• Machine Learning with Scikit-Learn

Future Improvements

• Hyperparameter Tuning
• Cross Validation
• Deployment using Streamlit or Flask
• Integration with Real-Time Car Market Data

How to Run

1. Clone this repository

git clone <repository-url>

2. Install required libraries

pip install pandas numpy matplotlib seaborn scikit-learn

3. Open the notebook

jupyter notebook car_price_prediction.ipynb

4. Run all cells sequentially

Author

Shreyansh Gupta B.Tech (CSDS) Galgotias College of Engineering & Technology