🚗 EV Remaining Range Prediction - Machine Learning Model

An end-to-end Machine Learning model trained on Project_Final_Dataset_CarID_Updated.xlsx that predicts an Electric Vehicle's (EV) remaining driving range in kilometers based on:

1. Manufacturer (Brand) & Model (with catalog lookup of battery capacity in kWh)
2. Current State of Charge / Battery Level (%)

---

📁 Project Structure

ML_EV/
├── Project_Final_Dataset_CarID_Updated.xlsx # Official EV Fleet Dataset (10,000 trips across 50 models)
├── EV_Range_Prediction_Model.ipynb          # Jupyter Notebook (EDA, data1 schema, ML modeling, evaluation)
├── train_model.py                           # Python ML training & export script
├── predict.py                               # CLI & Python inference utility
├── ev_range_model.pkl                       # Serialized trained model & vehicle catalog package
├── requirements.txt                         # Python dependencies
└── README.md                                # Project documentation

---

⚡ Quick Start

1. Run Inference / Predictions

Predict remaining range via command line:

# Predict for a specific brand and model
python predict.py --brand Tata --model "Punch EV" --battery 75

# List all supported EV manufacturers
python predict.py --list-brands

# List all models for a manufacturer
python predict.py --list-models Tata

2. Python Code Integration

import pickle

# Load the trained ML model package
with open("ev_range_model.pkl", "rb") as f:
    model_payload = pickle.load(f)

pipeline = model_payload["pipeline"]

# Run inference directly
# (Pass a DataFrame matching the data1 feature columns)

3. Re-train the Model

To re-train on Project_Final_Dataset_CarID_Updated.xlsx and export a fresh .pkl:

python train_model.py

4. Interactive Jupyter Notebook

Open and explore EV_Range_Prediction_Model.ipynb in VS Code / Antigravity using the Python 3.14.6 kernel.

---

📊 Feature Schema (data1)

Column	Description	Type
Range	Remaining Driving Range in km	Target ($y$)
Battery Level (%)	Current battery percentage	Feature ($X_1$)
Make	Vehicle Manufacturer	Feature ($X_2$)
Acceleration 0 - 100 km/h	0-100 km/h time (s)	Feature ($X_3$)
Top Speed	Top speed (km/h)	Feature ($X_4$)
ER	Full Electric Range (km)	Feature ($X_5$)
Total Power	Motor power (kW)	Feature ($X_6$)
Total Torque	Motor torque (Nm)	Feature ($X_7$)
Drive	Driver behaviour style	Feature ($X_8$)
BC	Battery capacity (kWh)	Feature ($X_9$)
Length, Width, Height, Wheelbase	Vehicle dimensions (mm)	Features ($X_{10\dots13}$)
Gross Vehicle Weight (GVWR)	Vehicle weight (kg)	Feature ($X_{14}$)
Max. Payload	Maximum payload capacity (kg)	Feature ($X_{15}$)
Cargo Volume	Cargo space (L)	Feature ($X_{16}$)

---

📈 Model Performance

Model Algorithm	$R^2$ Score	MAE (km)	RMSE (km)
Gradient Boosting Regressor	0.8960	23.25 km	29.93 km
Random Forest Regressor	0.8794	24.31 km	32.24 km
XGBoost Regressor	0.8751	24.71 km	32.81 km
Linear Regression	0.8593	27.61 km	34.82 km