Data science for energy engineers: the case of demand-side management

online course, 12-23 July 2021
EIT InnoEnergy

Data science for energy engineers: the case of demand-side management. Learn how to analyse, forecast, and optimise energy demand using Python.

What is this course about?  

The global energy system is undergoing a rapid transformation to meet climate goals. Digitization and data science have emerged as key enabling technologies to enable this transformation. Data science is increasingly being used by energy professionals to optimize energy flows, ensure decarbonization of the energy sector and minimize costs. This course provides a broad introduction to the many use cases of data in demand side management. Taking a hands-on approach, this course leads learners through the entire data science pipeline for a demand side management use case. Over the course, learners will learn how to forecast and optimize energy demand using a number of different tools in Python.

Is this course right for you?

This course is designed for energy domain professionals who are interested in learning more about the application of data science to real energy problems. As such, it is useful for not just students and researchers, but also for practitioners looking to incorporate data-driven decision-making into their skill set. 

By following this course, you will understand the end-to-end data pipeline, and how to make use of different data science tools to create additional value, ranging from analysis, modeling and optimization.

Why should you attend this course?

This course on data science for energy engineers has been designed based on industrial requirements, and feedback from over a hundred learners. It has been developed in collaboration with experts from leading European universities including KU Leuven, KTH, UPC and Grenoble INP. 

This course gives you the tools to: 

•    Get started on your data science journey for the energy domain

•    Level up your data-related programming skills with challenge-driven assignments 

•    Learn how to analyze your data in Python and present the results in a visually appealing way

•    Explore data-driven innovation to create a competitive advantage for your organization

•    Create forecasts with an emphasis on energy demand; however, with the tools you learn, you will be able to create just about any forecast in both the energy and the broader supply chain domain

•    Optimize energy demand using a battery-inverter system; however, the tools you learn in the course will enable you to run large-scale optimization on problems you face in your organization, whether they relate to energy demand or other elements in a broader supply chain

What will you learn?

This course empowers you to better understand energy data in the form of time series such as the ones frequently encountered in demand side management problems and create value from them. Specifically, learners will have the following trajectory during the course:

Session 1: 12 July 10:00-12:00 CEST

  • Introduce data science use cases in energy 
  • Introduce python programming with Jupyter notebooks
  • Provide a brief overview of helpful resources and existing datasets
  • Explain how to do exploratory data analysis in Python (for energy datasets)

ILO: Learners should be able to load various energy datasets, and visualize them for patterns as well as deal with issues such as missing values.

Session 2: 13 July 10:00-12:00 CEST

  • Introduce forecasting principles: what can and cannot be forecast
  • Provide overview of time series modelling techniques (benchmarks, exponential smoothing etc.)
  • Explain how to forecast using (Facebook) Prophet
  • Make forecasts for energy time series such as building energy demand

ILO: Learners should be able to understand core forecasting principles, and how to apply them to energy datasets.

Session 3: 14 July 10:00-12:00 CEST

  • Provide an overview of forecasting with machine learning techniques
  • Compare forecasts made with different techniques
  • Explain sources of uncertainty in predictions and some modern directions in forecasting research 
  • Introduce Kaggle competition (forecasting challenge)

ILO: Learners should be able to make more advanced forecasts with machine learning models, while also understanding their limitations.

Session 4: 15 July 10:00-12:00 CEST

  • Introduce different types of optimization problems in practice
  • Explain how to develop an optimal controller in Python to optimize for different objective functions (cost minimization, peak shaving)
  • Explain how to apply a number of different techniques to understand the different elements of an optimal control problem

ILO: Learners should be able to optimize the behavior of energy flexible resources, given a cost function.

Session 5: 16 July 10:00-12:00 CEST

  • Introduce advanced data science concepts for energy
  • Industry talks

ILO: Learners should be able to see the bigger picture surrounding individual algorithms in energy data science, and how they are applied in the energy industry.

Between sessions 5 and 6, learners will work on the competition / projects in the second week. 

An instructor will be on hand to help them. 

Session 6: 23 July 10:00-12:00 CEST

  • Overview of competition results and learner presentations
  • Wrap up and explain the challenges in creating end to end systems to deploy energy data science solutions

ILO: Learners should be able to present the results of their analysis in a manner accessible to both specialists and non-specialists.

How will you learn?

This course offers an immersive learning experience, where lectures explaining the theory of energy data science will be complemented with Jupyter notebooks in Python. The lectures will not just introduce classical concepts, they will also delve into some of the cutting-edge innovations in research in data science. The notebooks will enable learners to get hands-on experience with applying core concepts themselves. The course is designed in a way to encourage interaction, with learners able to ask instructors their questions during the sessions in addition to collaborating with each other on the online forum. The course also emphasizes the importance of applied data science, and learners will be expected to share their project results with the group.

Learners will be assessed based on a number of criterion: (1) minimum participation in the course, (2) in-class quizzes, (3) performance on a forecasting competition, and (4) presentation of the group project.

What will you achieve?

1.    A broad understanding of the many different use cases of data in the energy domain

2.    Concrete knowledge on how to analyze, forecast and optimize energy demand data in very general settings

3.    Certificate of attendance issued by EIT InnoEnergy and KU Leuven (3 ECTS).

Who will you learn from?

Dr. Ir. Hussain Kazmi (FWO research fellow, KU Leuven Postdoc researcher at KU Leuven)


  • 12-23 July 2021
  • online course
  • Language: English
  • Target audience: every energy professional who wants to learn about the application of data science to real energy problems


  • The course is full, but some spots remain for PhD researchers
  • Prerequisites:
    • Basic proficiency in Python (e.g., familiarity with control commands and loops, etc.). Experience with object-oriented programming is not required.
    • Understanding of core concepts in energy and power engineering (e.g., how energy is transmitted from the grid to the end-user, etc.).
  • Target audience: energy professionals

Ready to get started?

All information can be found on the website of InnoEnergy.