The Smart Energy Monitor is an end-to-end Machine Learning application designed to predict household energy consumption and optimize electricity bills.
Building upon my research experience at Chengdu University of Technology (CDUT), China, where I analyzed energy patterns of 2.2 million households, this project brings those insights to a practical, user-friendly dashboard tailored for the Pakistani context.
It combines Real-time Weather Data, Appliance Load Calculation, and AI Prediction to help users reduce their Carbon Footprint and Electricity Costs (PKR).
- 🇵🇰 Localized Context: Customized for Pakistani households (includes UPS, Water Motors, Iron, ACs).
- 💰 Bill Estimation: Real-time calculation of hourly and monthly costs in PKR (Rs.).
- 🤖 AI-Powered: Uses
Random Forest Regressorto predict Base Load based on weather conditions. - ⏱️ Real-Time Sync: Automatically fetches Pakistan Standard Time (PKT) for accurate simulation.
- 🌱 Green AI Audit: The model training process was audited using
CodeCarbonto ensure minimal energy consumption (0.12g CO2 footprint).
- Frontend: Streamlit (Custom CSS & Glassmorphism UI).
- Machine Learning: Scikit-Learn (Random Forest).
- Visualization: Plotly (Interactive Gauge Meters & Pie Charts).
- Sustainability: CodeCarbon (for tracking model efficiency).
- Weather Input: The user sets the current weather (Temperature, Humidity).
- Appliance Selection: Selects active devices (AC, Fans, Fridge, UPS charging).
- AI + Logic: The system combines the AI's "Base Load" prediction with the calculated "Appliance Load".
- Output: Provides a live dashboard showing total watts, estimated bill, and energy-saving tips.
# Clone the repository
git clone [https://github.com/your-username/smart-energy-monitor.git](https://github.com/your-username/smart-energy-monitor.git)
# Install dependencies
pip install -r requirements.txt
# Run the app
streamlit run app.py