Energy Trading Company: Predictive Analytics in Energy Market Trading

Energy Trading Company is a key player in trading electricity and electricity derivatives in Central Europe. It participates in major power exchanges in the region and actively manages the dispatch of physical electricity in power networks.

Energy Trading Company and KInIT cooperated within the Hopero project on the analysis of energy market data, aiming to understand the relationship between weather and electricity consumption, production and price. 

The goal of the project was to support decision making by predicting clearing prices and revealing characteristics of supply curves. Our research focused on two main directions. In the first phase, we used weather forecasts and predictions of production capacities to predict day-ahead market clearing prices. We applied statistical methods and  the latest machine learning and artificial intelligence techniques, evaluating our approaches on real market data.

Recognizing that supply curves also influence market clearing prices, the second part of the project focused on modeling supply curve characteristics. We used modern dimensionality reduction methods to identify a small number of key characteristics and developed a method to predict critical points of supply curves on a day-ahead basis.

Both Energy Trading Company and KInIT found the cooperation highly valuable and productive. The Energy Trading Company gained better insights into applications of machine learning and artificial intelligence for market data analysis, while KInIT developed a deeper understanding of energy markets and pricing mechanisms. Based on mutual satisfaction, we agreed to continue our collaboration in future projects.