🔋✨ Energy Consumption Prediction Solution

This repository contains a solution for predicting household energy consumption based on historical data and weather information.

📚 Table of Contents

1. 📖 Project Overview
2. 🔧 Prerequisites
3. 🚀 Quick Start
4. 📡 API Usage
5. 📊 Streamlit Dashboard
6. 🗂 Project Structure
7. 🤝 Contributing

📖 Project Overview

This project aims to predict the next value of the "active_power" column in a dataset of household energy consumption. The solution includes:

• 📈 Data preprocessing and feature engineering
• 🧠 Model training and evaluation
• 📊 A Streamlit dashboard for data visualization and model insights
• 🐳 Containerized model deployment
• 🌐 A Flask API for serving predictions
• ☁️ Cloud infrastructure design suggestion
• ⚙️ Considerations for production-scale operationalization

☁️ Cloud Infrastructure Schema

Our proposed cloud infrastructure leverages AWS services to create a scalable, maintainable, and secure solution.

For a detailed diagram and explanation of the infrastructure, please refer to the pdf file.

🔧 Prerequisites

• Python
• TensorFlow
• Streamlit
• Docker
• Flask

🚀 Setup and Installation

1. Clone this repository:

https://github.com/Ansem-chaieb/ve2max_assessment.git

cd Energy_Consumption_Prediction

2. Build and run the Docker container:

docker build -t energy-consumption-prediction .

docker run -p 5000:5000 energy-consumption-prediction

📡 API Usage

Access the Flask API:

• Via web browser or curl: http://localhost:5000
• Using Postman: Import the provided Postman collection:

{
    "datetime": "2024-01-05 23:59:00"
}

• Using Thunder Client VS Code extension:
  1. Install the Thunder Client extension in VS Code
  2. Click on the Thunder Client icon in the sidebar
  3. Create a new request to http://localhost:5000/predict
  4. Set the method to POST and add the JSON body as shown:

Note: To run the Flask API, please ensure that you have the AWS credentials properly configured in the .env file. If you wish to execute the API, I will need to send you the necessary credentials beforehand.

📊 Streamlit Dashboard

The Streamlit dashboard provides interactive visualizations of the dataset and model performance. It includes:

Access the Streamlit dashboard (EDA and model result) at: https://householdenergyconsumption.streamlit.app/

🗂 Project Structure

energy-consumption-prediction/
├── src/
│   ├── dataset.py        # Manages data loading, preprocessing, and train test split
│   ├── features.py       # Responsible for feature engineering and processing
│   ├── dashboard.py      # Contains code for creating the Streamlit dashboard
│   ├── plots.py          # Includes functions for generating various plots and visualizations
│   └── modeling/
│       ├── train.py      # Handles the training process 
│       └── predict.py    # Manages the prediction process using trained models
│
├── notebooks/
│   └── train_evaluate_models.ipynb    # Jupyter notebook for training and evaluating different models
├── docs/
│   └── Household_energy_consumption.pdf
├── Dockerfile
├── app.py                # Flask application
├── requirements.txt
└── README.md

🤝 Contributing Members

Data Scientist: Ansem Chaieb