Energy Consumption Prediction App

A predictive web application that uses a Random Forest Regression model to forecast hourly energy consumption. This app demonstrates how AI can assist the Department of Energy (DOE) by providing accurate, real-time energy demand predictions based on user inputs and historical data, in the repo is also included a presentation of the project as well as the jupyer notebook where the predictive data analysis was performed.

Features

• AI Model: Random Forest Regression model for hourly energy prediction.
• User Inputs: Customizable input fields for household size, home type, and appliance usage.
• Real-Time Prediction: Provides instant, hour-by-hour energy predictions to optimize resource planning.

Installation

1. Clone this repository:

   git clone https://github.com/your-repo-url.git
   cd your-repo

2. Install dependencies:

   pip install -r requirements.txt

3. Run the Flask application:

   python app.py

4. Access the app at http://127.0.0.1:5000 in your web browser.

Usage

1. Open the app in a browser.
2. Input household details, select home type and appliance usage level.
3. Click Get Prediction to view the forecasted energy consumption for the next hour.

Project Structure

• app.py: Main Flask application file.
• dataset/: Contains the energy dataset file.
  • PJME_hourly.csv: CSV file with historical energy consumption data.
• model/: Stores the machine learning model and scaler used in predictions.
  • energy_forecast_model.pkl: Pre-trained Random Forest Regression model.
  • scaler.pkl: Scaler for preprocessing input data.
• notebook/: Jupyter notebook(s) used for data analysis and model training.
  • PJM_Hourly_Energy_Consumption.ipynb: Notebook for data processing and model development.
• presentation/: Contains presentation files for project overview.
  • AI-Driven Predictive Energy Analytics.pptx: Presentation on AI applications for the DOE.
• templates/: HTML templates for the web interface.
  • index.html: Main HTML template for user input and prediction display.
• venv/: Python virtual environment for dependencies.
• README.md: Project documentation.
• requirements.txt: List of dependencies for the project.

Technologies Used

• Python (Flask): Backend application framework.
• HTML/CSS/JavaScript: Front-end interface.
• Pandas, Numpy, Matplotlib, Seaborn: For the data analysis and data visualization.
• Scikit-Learn: For the Random Forest Regression Machine Learning model.
• Joblib: Model serialization and deserialization.
• PDF: Presentation of the project.

Future Enhancements

• Live Data Integration: Pull real-time data from DOE sources for even more accurate forecasting.
• Enhanced Visualization: Add graphs to show historical vs. predicted energy usage patterns.