The next-generation scalable Forecasting solution: a part of our One Planet plan
Daniel Enthoven, Short-Term Trade Desk Scientist
These are exciting times for Forecaster and Data Scientist Daniel Enthoven of Eneco Energy Trade (EET). The renewable energy breakthrough has prompted a whole new dynamic in the energy market. In his previous blog, Daniel discussed the inviting challenges of his work at the Short Term Trade Desk. Today, he goes into more detail about what rising to those challenges entails.
Energy Trading
My Data Science Team and I specialise in all facets of forecasting, which is an increasingly important task in the light of Eneco’s big ambition to become climate neutral by 2035. As part of efforts to achieve this ambition, we’re rapidly expanding our sustainable asset portfolio of solar and wind parks.
Our energy traders deal in energy produced by third parties as well as assets operated by Eneco. They need to know as accurately as possible how much energy will be produced in the short and long term, as this information is a key input for risk management and the further development of an optimum trading strategy that balances supply and demand. Providing that information is where my colleagues and I come in.
Our energy traders deal in energy produced by third parties as well as assets operated by Eneco. They need to know as accurately as possible how much energy will be produced in the short and long term, as this information is a key input for risk management and the further development of an optimum trading strategy that balances supply and demand. Providing that information is where my colleagues and I come in.
Build valuable models
The forecast team has focused its attention in the recent past on building a new generation of software and models. The new software is a scalable system with which forecasts can be made for 10 assets, but also for say 10,000. It runs on the Eneco Data Platform, a powerful infrastructure that can easily handle the new software and huge amounts of data. The models take advantage of machine learning, yielding more precise predictability, better insights, and minimal risk.
Within this environment, we develop forecasting products with which we provide traders with valuable information.
Within this environment, we develop forecasting products with which we provide traders with valuable information.
Models for power generation from solar and wind, as I just mentioned, or other related products: models the predictability of which has a direct impact on market prices.
We took these models from idea to implementation in a very brief timeframe. That’s quite extraordinary, but at the same time typical of how we work at EET. If one of us has a well-thought-out idea for a product that will add value, we set about making it a reality together, as a team. We immerse ourselves in the algorithms and business logic and do everything we can to quickly get the idea into the production environment.
We took these models from idea to implementation in a very brief timeframe. That’s quite extraordinary, but at the same time typical of how we work at EET. If one of us has a well-thought-out idea for a product that will add value, we set about making it a reality together, as a team. We immerse ourselves in the algorithms and business logic and do everything we can to quickly get the idea into the production environment.
Work on more accurate output and input
When forecast products go live, it doesn’t mean they’re finished. They’re organic in that sense, requiring constant attention. That’s because in practice, things sometimes turn out to be different from what was envisaged on the drawing table, or because external circumstances have changed. Whatever the reason, we find the cause and devise a solution as a team, as we seek always to fine-tune our products in response to market developments. We discuss what, at any given time, will add the greatest value and then pursue it.
What helps, is that the Short-Term Trade Desk consists of a group of very diverse forecasters with different backgrounds. One person specialises in processes, another in finance, and yet another knows all about data science.
What helps, is that the Short-Term Trade Desk consists of a group of very diverse forecasters with different backgrounds. One person specialises in processes, another in finance, and yet another knows all about data science.
Everyone brings their own unique qualities to the table and provides the team as a whole with correspondingly unique feedback. Together, we make sure the end product sees the light of day.
But there’s more to this game than programming alone. However complex a model may be, it’s only ever as good as the data you feed it. To tackle this problem, we spend a lot of time increasing data veracity and decreasing variables such as weather and availability bias. That way, we leverage the strength of our models to the max and maintain sustained forecast quality improvements.
But there’s more to this game than programming alone. However complex a model may be, it’s only ever as good as the data you feed it. To tackle this problem, we spend a lot of time increasing data veracity and decreasing variables such as weather and availability bias. That way, we leverage the strength of our models to the max and maintain sustained forecast quality improvements.
Contribute to a better future
I really enjoy the interplay and overlap between strategic product development and operational maintenance. It makes my job exciting and means that I get to work in close cooperation with colleagues in all kinds of different positions. We’re always in direct contact with the intraday traders, for instance, with whom we maintain a short feedback loop. Such collaboration allows me to increase my knowledge of the market and our organisation, which in turn enables me to identify opportunities for new forecasting products to provide even more effective support to the trading team.
Working together with my colleagues also sharpens my awareness of our role in the grander scheme of things. If we notice, for instance, that we’re not receiving sensor data from a solar park, or that it’s producing less electricity than normal, then we get in touch with the relevant departments so that they can investigate and resolve the issue. And whenever a new park comes online somewhere, we add it to the energy portfolio. That’s how I contribute directly to a healthy Eneco – and a climate-neutral future.
Working together with my colleagues also sharpens my awareness of our role in the grander scheme of things. If we notice, for instance, that we’re not receiving sensor data from a solar park, or that it’s producing less electricity than normal, then we get in touch with the relevant departments so that they can investigate and resolve the issue. And whenever a new park comes online somewhere, we add it to the energy portfolio. That’s how I contribute directly to a healthy Eneco – and a climate-neutral future.
Find your new digital or tech job at Eneco:
Would you like to receive our newest job vacancies?
This website uses cookies
For the proper functioning and anonymous analysis of our website, we place necessary and functional cookies,
which have no consequences for your privacy.
We use more cookies, for example to make our website more relevant to you,
to make it possible to share content via social media and to show you relevant advertisements on third-party websites.
These cookies may collect data outside of our website. By clicking "Accept all cookies" By clicking you agree to the placing of these cookies.
You can find more information in our cookie policy and our privacy statement.