Historical Trends and Predicting Future Energy Efficiency and Demand in Clean Energy Transitions Using Machine Learning and Deep Learning Techniques between 1900 to 2074
Carlos Vergara Gámez
E-mail: [email protected]
🗺 In this project, you can generate a fully interactive world map and Europe map to explore renewable energy usage in an engaging way!
The Sustainable Energy Analysis Platform (SEAP) is my final project from the Data Analytics Bootcamp at Ironhack. The project focuses on predicting renewable energy generation trends for solar, wind, hydroelectric, and biofuel sources. By leveraging Machine Learning and Deep Learning, the goal is to assess energy efficiency and demand during the transition to clean energy sources.
- 🌲 Multi-Output Random Forest for Multivariate Time Series Forecasting Optimized with Grid Search (MO-RF-MTSF-GS).
- 🧠 Recurrent Neural Network with Multilayer Perceptron for Multivariate Time Series Forecasting (RNN-MLP-MTSF).
- Predict future energy production trends for renewable sources.
- Analyze energy demand and efficiency in clean energy transitions.
- Provide interactive visualizations for data exploration and predictions.
- Develop advanced AI models for real-time energy insights.
This project leverages a variety of tools:
- Python, Pandas, NumPy for data processing.
- Matplotlib, Seaborn, Plotly for visualizations.
- Scikit-Learn for Machine Learning models.
- TensorFlow/Keras for Deep Learning.
- Geopandas and Folium for spatial data analysis.
The dataset used in this project comes from:
Ritchie, H., Rosado, P., & Roser, M. (2023). Energy. Published at OurWorldinData.org. Link
- 📂 SEAP_01_EDA_map_and_cleaning.ipynb: Exploratory Data Analysis and cleaning, including map visualizations.
- 📂 SEAP_02_Machine_Learning_and_Deep_Learning.ipynb: Implementation of Machine Learning and Deep Learning models for energy prediction.
-
Clone the repository and set up the environment:
conda env create -f ml-dp.yml conda activate ml-dp
-
Download necessary folders:
dataandshapefiles. -
Open and run the notebooks:
SEAP_01_EDA_map_and_cleaning.ipynbSEAP_02_Machine_Learning_and_Deep_Learning.ipynb
-
Generated models and maps will be saved in the
final_modelsandmapsfolders.
-
Global Renewable Energy Trends:
The graph shows a significant increase in renewable electricity production, with solar and wind energy growing over time, while hydroelectric power remains dominant. It also shows an exponential increase in electricity demand since 2000, making renewable energy sources currently insufficient to fully replace non-renewable energy.
-
Energy Demand vs. Supply:
The chart compares predicted and actual electricity demand, highlighting the model's accuracy.
Looking ahead, my goal is to develop an AI system capable of managing larger datasets through Deep Learning techniques. This AI will provide real-time predictions and enable interactive communication, allowing users to explore energy-related challenges and propose sustainable solutions.
I would like to express my heartfelt gratitude to Ironhack Bootcamp for all the invaluable knowledge and skills I've acquired throughout this journey. The experience has not only been educational but also incredibly enriching on a personal level.
A special thanks to my amazing mentors, Santiago, Antonio, and Nicolás, whose guidance and support have been instrumental in my learning process. Your insights and encouragement have motivated me to push my limits and strive for excellence.
I also want to extend my appreciation to all my fellow classmates. Thank you for the countless moments of laughter, collaboration, and hard work we shared together. These memories will stay with me as a testament to our collective journey.
Thank you all for being a part of this transformative experience!
Carlos Vergara Gámez