AICTE – Edunet Foundation Virtual Internship
Theme: Electrical Vehicles and Energy
Internship Partner: Edunet Foundation| Shell
Electric Vehicles (EVs) are revolutionizing global transportation by offering a cleaner, more efficient alternative to traditional fuel-based vehicles.
As the world moves toward sustainability, understanding what drives the price of an electric vehicle is vital.
This project focuses on using Machine Learning to predict the price of an EV based on its technical specifications — such as range, efficiency, acceleration, and top speed.
The insights gained can assist both manufacturers and customers in making better, data-driven decisions.
Develop a machine learning model that predicts the price of electric vehicles using multiple performance and design parameters.
- Collect and preprocess an EV dataset from Kaggle.
- Perform exploratory data analysis (EDA) and feature engineering.
- Train regression models to predict prices accurately.
- Compare model performance using error metrics.
- Present results through visualizations and documentation.
Dataset Name: Vehicle Dataset 2024
Source: Kaggle – Vehicle Dataset 2024
File Name: car_data_clean.csv
| Feature | Description |
|---|---|
| Range (km) | Distance covered on a full charge |
| Top Speed (km/h) | Maximum achievable speed |
| Efficiency (Wh/km) | Energy consumed per kilometer |
| Acceleration (0–100 km/h) | Time to reach 100 km/h |
| Seats | Seating capacity |
| Price (USD) | Market price of the EV |
- Data Cleaning – Removed missing and duplicate values.
- Feature Selection – Identified the most influential features.
- Encoding – Converted categorical variables into numerical form.
- Scaling – Normalized numeric columns for uniformity.
- Splitting – Divided the dataset into 80% training and 20% testing sets.
- Programming Language: Python 🐍
- Environment: Google Colab / Jupyter Notebook
- Libraries Used:
- pandas – Data handling
- numpy – Numerical operations
- matplotlib / seaborn – Visualization
- scikit-learn – Machine Learning models
- joblib – Model persistence
- Linear Regression – Baseline model
- Decision Tree Regressor – Handles non-linearity
- Random Forest Regressor – Ensemble method for better accuracy
- Mean Absolute Error (MAE)
- Root Mean Squared Error (RMSE)
- R² Score
Result:
The Random Forest Regressor achieved the best accuracy and the lowest prediction error, proving to be the most reliable model for EV price estimation.
- EVs with longer range and better efficiency are more expensive.
- Acceleration and top speed significantly influence the final price.
- Energy efficiency is a key factor determining cost-performance ratio.
- Gained practical knowledge in data preprocessing and visualization.
- Learned to implement and evaluate regression algorithms.
- Understood feature correlation and its impact on predictions.
- Practiced project documentation and GitHub publishing.
- Integrate Deep Learning (Neural Networks) for better accuracy.
- Create a Streamlit-based web interface for live predictions.
- Add brand-wise comparisons and battery performance analytics.
- Expand dataset for real-world applicability.
⚡ “Powering Tomorrow’s Electric Revolution”
Predicting the Price of Electric Vehicles through Data Intelligence
This project was developed as part of the AICTE – Edunet Foundation Virtual Internship, under the theme:
*Electrical Vehicles and Energy, in partnership with *Shell.
Special thanks to the mentors and coordinators for their valuable guidance and training sessions.
Theme: Electrical Vehicles & Energy
Partner: Edunet Foundation | Shell
Predicting the Price of Electric Vehicles through Data Intelligence
Electric Vehicles (EVs) are transforming transportation by offering cleaner and more efficient alternatives. Understanding the factors that influence EV prices helps customers, researchers, and manufacturers make informed decisions.
This project uses Machine Learning to predict the price of an electric vehicle using features like range, speed, efficiency, acceleration, and more.
- GitHub repository created and structured
- Datasets added (cleaned CSV files)
- Initial README drafted
- Performed basic data checks: missing values, duplicates, outliers
- Analyzed feature distributions and dataset structure
Week 1 Improvements:
- Organized folders and cleaned the dataset
- Added improved README structure
- Added output screenshots to documentation
- Loaded dataset in Google Colab
- Performed data preprocessing:
- Cleaning
- Feature Selection
- Encoding
- Scaling
- Trained the following models:
- Linear Regression
- Decision Tree Regressor
- Random Forest Regressor (best performance)
- Evaluated using MAE, RMSE, and R² score
- Exported model and scaler for next weeks
Files Added in Week 2:
Week2_Model_Training.ipynb– ML workflow notebookev_price_model.pkl– Trained Random Forest modelscaler.pkl– Scaler used during training
Week 2 Insights:
- Higher range EVs tend to have higher prices
- Efficiency and performance influence pricing significantly
- Random Forest provided the most reliable predictions
- Week 3 – Model Integration & Python Front-End Development
- Integrated the trained Random Forest model and scaler into the Python-based front-end.
- Created a complete front-end interface using Python files in VS Code.
- Implemented user input handling for EV specifications (battery, brand, range, power, etc.).
- Performed all preprocessing stages inside the Python front-end:
- Encoding
- Scaling
- Implemented the price prediction logic using the imported ML model.
- Displayed predictions and results in the Python front-end interface.
- Organized and cleaned the entire project structure for final submission.
- Files Added in Week 3:
frontend_react website.zip– Python front-end interface
This is the Website's Frontend attached to the final project and ml model zipped together
- Week 3 Insights:
- Python-based front-end successfully connected with ML model for real-time predictions.
- Fully functional EV Price Prediction workflow implemented without using Streamlit.
- Final project optimized for LMS submission with a clean and lightweight structure.
Dataset Name: Vehicle Dataset 2024
Source: Kaggle
File Used: car_data_clean.csv
| Feature | Description |
|---|---|
| Range (km) | Distance per full charge |
| Top Speed (km/h) | Maximum specified speed |
| Efficiency (Wh/km) | Energy consumption per km |
| Acceleration (0–100 km/h) | Time taken to reach 100 km/h |
| Seats | Seating capacity |
| Price (USD) | Target variable |
- Python
- Google Colab
- Pandas, NumPy
- Matplotlib, Seaborn
- Scikit-learn
- Joblib
- Linear Regression (baseline)
- Decision Tree Regressor
- Random Forest Regressor – Best Model
- Mean Absolute Error (MAE)
- Root Mean Squared Error (RMSE)
- R² Score
- Random Forest delivered highest accuracy
- EV range & efficiency strongly influence pricing
- Model exported for chatbot and Streamlit app
EV-Price-Prediction/ │── README.md │── ML Training Model.ipynb │── ev_price_model.pkl │── scaler.pkl │── ev_sales_adoption.csv │── CarPrice_Assignment.csv │── frontend_react website.zip
- Data preprocessing & visualization
- Regression model training & evaluation
- Understanding feature correlations
- Model saving & deployment preparation
- GitHub documentation