README.md

---
license: mit
tags:
  - regression
  - electricity
  - ev-charging
  - price-prediction
  - random-forest
  - scikit-learn
  - tabular-data
language:
  - en
library_name: scikit-learn
datasets:
  - processed_electric_price_filled.csv
metrics:
  - r2
  - mse
  - mape
model-index:
  - name: Electricity Price Predictor
    results: []
---


# 🔋 Electricity Price Predictor (Random Forest Regression)

This is a custom regression model trained to predict electricity prices ($/kWh) in California, based on a variety of grid-level and environmental features such as EV charging demand, solar/wind production, carbon emissions, and storage indicators.

The model is trained using `RandomForestRegressor` from `scikit-learn`, with 24 engineered features and a structured tabular dataset. This project is intended to support intelligent energy systems, such as EV charging optimization, energy scheduling, or smart grid simulation.

---

## 📌 Model Details

- 📈 **Model**: RandomForestRegressor (n_estimators=200)
- 🧠 **Framework**: scikit-learn
- 🧾 **Input Features**: 24 numerical values (see full list below)
- 🎯 **Target Variable**: Electricity Price ($/kWh)
- 🗃️ **Data**: Structured time-series dataset with hourly EV/grid info
- 🧪 **Evaluation**:
  - MSE: *e.g., 0.0023*
  - R²: *e.g., 0.89*
  - MAPE: *e.g., 6.5%*

---

## 🔢 Input Features

The model expects a list of 24 numeric features:

```text
['Year', 'Month', 'Day', 'DayOfWeek', 'Hour',
 'EV Charging Demand (kW)', 'Solar Energy Production (kW)', 'Wind Energy Production (kW)',
 'Battery Storage (kWh)', 'Charging Station Capacity (kW)', 'EV Charging Efficiency (%)',
 'Number of EVs Charging', 'Peak Demand (kW)', 'Renewable Energy Usage (%)',
 'Grid Stability Index', 'Carbon Emissions (kgCO2/kWh)', 'Power Outages (hours)',
 'Energy Savings ($)', 'Total_Renewable_Energy_Production', 'Effective_Charging_Capacity',
 'Adjusted_Charging_Demand', 'Net_Energy_Cost', 'Carbon_Footprint_Reduction',
 'Renewable_Energy_Efficiency']

```
---
---

## 🧪 Usage Example

### 🔹 Option 1: Load and use the model directly

```python
import joblib
import numpy as np

# Load trained model
model = joblib.load("random_forest_model.pkl")

# Sample input (replace with actual values)
features = [0.5] * 24

# Make prediction
price = model.predict(np.array(features).reshape(1, -1))[0]
print(f"Predicted Electricity Price: ${price:.4f}")
```

---

### 🔹 Option 2: Use helper function in `predict.py`

```python
from predict import predict

features = [0.5] * 24
result = predict(features)
print(f"Predicted Price: ${result:.4f}")
```

---

### 🔹 Option 3: Try it online (Gradio Web Demo)

If deployed, you can try it here:  
[👉 Live Demo on Spaces](https://huggingface.co/spaces/YOUR_USERNAME/electricity-price-predictor-demo)

---

## 📊 Sample Dataset

This repository includes a sample dataset: `processed_electric_price_filled.csv`.  
It contains hourly records of EV charging demand, solar/wind energy production, grid stability, and electricity prices.

### Load and explore:

```python
import pandas as pd

df = pd.read_csv("processed_electric_price_filled.csv")
print(df.head())
```

---

## 📁 Files Included

| File                                | Description                                      |
|-------------------------------------|--------------------------------------------------|
| `random_forest_model.pkl`           | Trained RandomForestRegressor model             |
| `predict.py`                        | Python function to load and run predictions     |
| `app.py` (optional)                 | Gradio-based interactive demo                   |
| `requirements.txt`                 | Python dependencies                             |
| `processed_electric_price_filled.csv` | Training/test dataset                         |
| `README.md`                         | This documentation                              |

---

## 👨‍💻 Author

**bajiang(Georgia)**  

---

## 📄 License

MIT License – You are free to use, modify, and distribute this project with proper attribution.