Forecasting Future Energy Prices: An AI-Enabled Engine To Power AP’s Price Planning
By UBP Data Science and AI Group and AboitizPower ETS Team
We live in a world where smart devices are everywhere. They are “smart” mainly because they possess what we call artificial intelligence or AI. AI has been helping us detect spam emails for a while but it is only recently that we have been seeing breakthroughs covering various applications in different fields.
AI can now beat professionals in the field of radiology, particularly in brain tumor diagnosis, with record-breaking accuracy and speed. In economic development, AI can directly measure the poverty of inaccessible places through satellite images, which is helping policymakers decide where to prioritize their projects. In e-commerce, we’ve experienced how online shopping sites use AI in profiling digital footprints or the trail of data one leaves on their platform and recommend products tailored to an individual’s needs. There are certainly many more applications, most of which are present in our smartphones, smartwatches, and other smart devices.
With so much information available to it, AI is like our natural intelligence, learning through examples and experiences. For example, some radiologists learned to detect cancer from several samples of images of cancer tumors from their education. At the same time, others knew more from their own separate set of experiences. These events are similar to how AI learns too, by going through each example’s specifics or reviewing its previous mistakes. In fact, in the context of AI, these examples are what we refer to as data; and the science of learning from the data defines the field of data science, and those in this profession are called data scientists. Therefore, we can say that AI and data science are related, with data science being the factory that engineers the artificial intelligence of smart devices or tools.
The examples mentioned so far make a clear case as to why companies must take advantage of the capabilities of AI. Recognizing this, the Aboitiz Group is investing heavily in making data science and AI part of the decision-making process across SBUs. Indeed, there are many opportunities to take advantage of various available datasets to help target potential customers and improve current customer experiences. For example, in power, the goal is to innovate in delivering reliable and sustainable electricity at reasonable and competitive prices. AboitizPower, being the top power-generating company in the country, is embracing AI and adapting to today’s fast-changing technologies. The company teamed up with UnionBank in optimizing its energy trading activities by introducing AI to supplement the traders.
The project aims to produce AI-enabled tools for all forecasting activities, which include energy demand forecasting and price forecasting*. This will not only help AboitizPower anticipate unexpected economic fluctuations brought by sudden effects of external factors, such as pandemic, but also help the company stay true to its commitment to provide secure, reliable, and affordable electricity, energizing the dreams and hopes of the Filipino people for a better future.
AboitizPower, like other power generation companies, comes up with optimal prices—albeit manually—based on the traders’ expert assessment in order to remain competitive. With the growing renewable energy sector coupled with unexpected external factors, energy prices have become even more volatile for traders to forecast. On top of that, traders need to adapt to the soon-to-be-implemented five-minute trading market, which will translate into increased forecast points to fill in. Low accuracy in the forecast due to these variable factors can have considerable cost implications for the company, a loss in opportunity.
In exploring how to turn its “data as the new currency”, AboitizPower teamed up with UnionBank’s Data Science and AI Group to build an AI tool for its energy trading activities. The product is designed to help traders automate and optimize the central part of their routine processes and tasks, which saves time and resources. Tests on the first forecasting tool have yielded positive results that have even outperformed current methods. This encouraged end-users to expand their project and automate other processes as well, which will lead to further savings. Currently, this first forecasting tool is undergoing its second phase of improvement, which will incorporate recent advancements in energy forecasting activities.
With Aboitiz’s digital DNA built on agile and DevOps principles, the project teams are concurrently developing the second forecasting tool to maximize time. These tools aim to equip traders with technologies that can help them better position their market pricing.
Lastly, the teams are looking into expanding the applications of data science and AI in the energy trading field by exploring opportunities that would look deeper into the use cases of customer profiling, high-frequency forecasting, and grid optimizations. An exciting future indeed built on data-driven solutions!
Equipping A-People On Data Science & AI Basics
To demystify analytics and encourage A-People to explore professional development opportunities in the area, Aboitiz Academy has embarked on a series of online brown bag sessions conducted by UnionBank Senior Adviser for Data & AI David Hardoon.
David and his team are working with various centers, groups, and units to reinforce data infrastructure and governance, behavior modelling, machine learning, and AI capabilities as well as applications both at UnionBank and across the Aboitiz Group.
What do you think?