⚡ Energy Consumption Prediction using ANN

A machine learning project that predicts the heating load of buildings using an Artificial Neural Network (ANN), built with TensorFlow/Keras and deployed as a live interactive dashboard using Streamlit.

---

📌 Project Overview

Buildings consume a significant amount of energy for heating and cooling. Predicting energy consumption in advance helps in:

• Reducing electricity costs
• Efficient energy planning
• Building smarter, eco-friendly systems

This project trains an ANN on the UCI Energy Efficiency Dataset to predict heating load based on 8 building parameters. Users can interact with a live dashboard and get instant predictions by adjusting input sliders.

---

🎯 Features

• ✅ Data preprocessing with MinMax Normalization
• ✅ ANN model with 3 hidden layers + Dropout (anti-overfitting)
• ✅ Early stopping during training
• ✅ Model evaluation with MAE, RMSE, R² Score
• ✅ Training loss curve visualization
• ✅ Actual vs Predicted chart
• ✅ Live interactive Streamlit dashboard
• ✅ Real-time prediction on slider change

---

🗂️ Project Structure

energy-consumption-prediction/
│
├── ENB2012_data.xlsx       # UCI Energy Efficiency Dataset
├── train_model.py          # Train ANN and save model
├── dashboard.py            # Streamlit live dashboard
├── .gitignore
├── LICENSE
└── README.md

After running train_model.py, these files are generated locally:

• energy_model.keras — saved trained model
• scaler.pkl — saved MinMax scaler
• loss_curve.png — training vs validation loss chart
• actual_vs_predicted.png — prediction comparison chart

---

📊 Dataset

Source: UCI Machine Learning Repository — Energy Efficiency Dataset

Column	Feature	Description
X1	Relative Compactness	Shape of building
X2	Surface Area	Total surface in m²
X3	Wall Area	Wall surface in m²
X4	Roof Area	Roof surface in m²
X5	Overall Height	Height of building (m)
X6	Orientation	Direction building faces
X7	Glazing Area	Window/glass area ratio
X8	Glazing Area Distribution	Window placement
Y1	Heating Load	Target variable (kWh)
Y2	Cooling Load	Not used in this project

• Rows: 768
• Missing values: None
• Format: Excel (.xlsx)

---

🧠 ANN Architecture

Input Layer     →  8 neurons  (X1 to X8)
Hidden Layer 1  →  64 neurons (ReLU) + Dropout(0.2)
Hidden Layer 2  →  32 neurons (ReLU) + Dropout(0.2)
Hidden Layer 3  →  16 neurons (ReLU)
Output Layer    →  1 neuron   (Predicted Heating Load)

• Optimizer: Adam
• Loss Function: Mean Squared Error (MSE)
• Epochs: 200 (with Early Stopping)
• Batch Size: 32
• Train/Test Split: 80/20

---

⚙️ Installation & Setup

1. Clone the repository

git clone https://github.com/subham23s/energy-consumption-prediction.git
cd energy-consumption-prediction

2. Create and activate virtual environment

python -m venv venv

# Windows
venv\Scripts\activate

# Mac/Linux
source venv/bin/activate

3. Install required libraries

pip install tensorflow pandas numpy matplotlib scikit-learn streamlit openpyxl

---

🚀 How to Run

Step 1 — Train the model (run once)

python train_model.py

This will:

• Load and preprocess the dataset
• Train the ANN model
• Save energy_model.keras and scaler.pkl
• Generate loss and prediction charts
• Print MAE, RMSE, R² scores

Step 2 — Launch the live dashboard

streamlit run dashboard.py

Browser opens at http://localhost:8501 automatically.

---

📈 Model Performance

Metric	Description
MAE	Mean Absolute Error
RMSE	Root Mean Squared Error
R²	Coefficient of Determination (closer to 1 = better)

---

🖥️ Dashboard Preview

The live dashboard allows users to:

• Adjust 8 building parameter sliders
• Get instant prediction of heating load
• View normalized input feature chart
• See training loss and prediction charts

---

🛠️ Tech Stack

Tool	Purpose
Python 3.x	Core language
TensorFlow / Keras	ANN model
Scikit-learn	Preprocessing & metrics
Pandas & NumPy	Data handling
Matplotlib	Visualization
Streamlit	Live dashboard

---

👨‍💻 Author

Subham — AI/ML Engineering Student
GitHub: @subham23s

---

📄 License

This project is licensed under the MIT License.