The quality of the forecasts exceeded our expectations. Now we can plan our customers' energy requirements much better and manage energy use more efficiently than before.
Andreas Schmallegger, Abteilungsleiter Energiehandel, Energie Steiermark
Artificial Intelligence, Big Data, and Data Science make it possible: Austria’s fourth-largest energy service provider predicts the energy demand of its customers down to the quarter-hour for one month in advance, ensuring a significantly improved sustainable control of energy usage.
The launch of our project was in the middle of the Corona pandemic and we implemented almost the entire project remotely due to lockdowns. But even this hurdle we have mastered together through the great cooperation with the professional team of ACP CUBIDO.
Andreas Antensteiner, Teamleiter BI, Energie Steiermark
Energie Steiermark has been optimizing the forecast of energy consumption for a long time. Given the switch to digital smart meters and the enormous growth in data that this entails, it was decided to fundamentally modernize the BI landscape.
A solution was sought that would, on the one hand, provide a complete overview of electricity consumption and, on the other, deliver accurate energy demand forecasts for the more than 600,000 customers. Smart energy management that enables the targeted purchase of energy and sustainably reduces energy procurement costs. Technical requirements included careful separation of individual clients and easily scalable performance for optimized processing of many billions of data records. Within the scope of a tender, Energie Steiermark decided in favor of ACP CUBIDO and an innovative concept for the development of a Modern Data Warehouse.
Efficient data analysis with Microsoft Azure, Spark and Power BI
The foundation for this is an enterprise-wide data platform that was implemented using Microsoft Azure and Azure Synapse Analytics. It combines a powerful data warehouse under one roof and enables the processing of massive amounts of data with Spark. Machine learning processes, for example, are handled via this platform. In the course of data preparation, many billions of rows are processed and are available for daily queries with low response times. For comparison: In the past, such data evaluation was only carried out if required after time-consuming manual data preparation.
Data analysis is performed using Power BI, which does not require any specialized IT knowledge for report generation. At the same time, experienced data scientists now have access to new models and parameters that allow them to work more effectively and create better forecasts. In addition to the Azure Cloud, the new solution includes different database technologies and data domains that are provided on-premises via source systems. This serves to comply with data protection and regulatory requirements for the protection of sensitive customer data.
The path to smart data utilization
The specific requirements of Energie Steiermark were implemented in two successive project phases. First, a seamless review of past events was created, laying the groundwork for an insightful look into the future:
The setup of the analysis platform in the Microsoft Azure Cloud forms the structural basis for optimal data security as well as the strict separation of different clients. In addition, it offers a data-based view of energy consumption, thus already providing the source material for the qualified forecasts in the next step.
For a meaningful forecast, Energie Steiermark needs reliable data at 15-minute intervals for the past few years. This is achieved through the analysis of individual customer load profiles from the past and filling technical gaps. Based on this forecast foundation, electricity consumption is predicted with high accuracy using machine learning models. As a result, the required energy resources can now be planned in a targeted and significantly more cost-efficient manner.
Decide professionally, expand flexibly
The modern data warehouse and centralized data preparation ensure data consistency across all departments, thereby enhancing the quality of strategic decision-making across functional areas. At the same time, the new structure allows future new customers and churn to be taken into account for the preparation of reliable forecasts.
In addition, the new infrastructure was designed in such a way that it can be flexibly further developed and adapted to future requirements. The implementation of additional analytics use cases is therefore just as possible as the connection of new clients or technological enhancements in the areas of data science and machine learning. The close cooperation between the technical specialists of ACP CUBIDO and the experts of Energie Steiermark has created the necessary basis for driving forward innovative adaptations and developing new use cases.
ACP CUBIDO is a specialist in the optimization and automation of business processes, the development of individual software solutions as well as Industrial IoT and Data Management. Using innovative technologies such as artificial intelligence, machine learning and data science, the approximately 50 experts at ACP CUBIDO support their customers in the collection, analysis and processing of data of all kinds. In this way, valuable information can be extracted from company data.
Text translated from the original –ACP CUBIDO