27 July 2021 | Reading 10 mins.

Enabling more open science in the energy sector

Share

Jakob Rager
Director of CREM
Eric Wilczynski
Researcher at EURAC
Alessia Bardi
Researcher at CNR-ISTI/OpenAIRE
Giuseppe Peronato
Researcher at Idiap Research Institute
Giulia Conforto
Researcher at ethink
Simon Pezzutto
Senior Researcher, Eurac Research

Jakob Rager, Director of CREM

Eric Wilczynski, Researcher at EURAC

Alessia Bardi, Researcher at CNR-ISTI/OpenAIRE

Giuseppe Peronato, Researcher at Idiap Research Institute

Giulia Conforto, Researcher at ethink

Simon Pezzutto, Senior Researcher, Eurac Research

Accelerating the energy transition is an urgent priority for Europe, with scientists indicating that keeping global temperature rise to under 2°C is becoming increasingly unlikely. A series of necessary, but ambitious environmental targets have been set by the EU Green Deal, yet these mean reforms are needed across the energy sector, with further research required for us to better understand how to facilitate this change. Therefore, for Europe to achieve these goals, accurate, open-source data is a priority to provide scientists and energy experts with the information needed to protect our planet’s future and accelerate the energy transition.

Open Science challenges in the data energy sector

The internet has revolutionized our lives, giving us instant access to vast amounts of information. In the research field, this has been advantageous to many researchers who are able to share their work with a wider community and get access to more data by themselves. Gradually scientific publication platforms have been established as centralized databases hosting the published work of researchers around the world. This is how the concept of ‘Open Science’ started to grow – open access to research and information available to all to facilitate faster, more effective research. In an attempt to boost European research, the European Commission has made the ‘Open Science’ concept one of its most important principles in its communication.

According to Open Science (OS) principles, all scientific products generated by research activities (e.g. scientific literature, research data, software, experiments) should be as open access as possible, made available as soon as possible, and under terms that enable reuse, redistribution and reproduction of the research and its underlying data and methods. Therefore, scientists should follow OS publishing practices, but their adoption is often hindered by cultural and technical barriers.

Barriers to Open Science

  • Article-centric publishing: many research communities still show little interest in publishing research products beyond the scientific article; software publishing is still largely overlooked, and digital experiment publishing (e.g., methods, workflows, research objects) is rare
  • Article-centric scientific reward practices: little support for scientific reward systems that benefit diverse research products
  • Fragmentation of research products: products are often published into community-independent data sources, where they lose their ‘community flavor’; identifying the collection of products of one community across such sources is in many cases not straightforward or requires high manual/technical costs
  • Static publishing: publishing occurs at the end of the research activity, as an act of “freezing” the products by deposition onto scholarly communication data sources; any event taking place after publishing to the products (e.g. citation, new version, usage in another experiment) is not dynamically reflected/materialized in the data sources with proper semantic links.

Overcoming barriers to Open Science

Several initiatives are taking actions to overcome the barriers to Open Science by developing tools, best practices and training for Open Science. OpenAIRE is one of such initiatives funded by the EC, supporting research communities with the OpenAIRE CONNECT platform. It engages with stakeholders and encourages Open Science practices while monitoring the uptake of Open Science principles in the research community. At the heart of the CONNECT platform is the OpenAIRE Research Graph, an open resource that includes metadata and links between 150 million scientific products (literature, datasets, software, and other research products), as well as information on research organizations, funders, projects, communities, and research infrastructures. The result is an open, trusted, public resource enabling the exploration of the scholarly communication landscape.

More specifically for the energy research community, the Horizon 2020 project EnerMaps, has developed a gateway to act as a single entry-point to scientific products of any type produced in the energy field, as well as tools that support and monitor the uptake of Open Science practices in the community.

This tool for interdisciplinary collaborations combined with the latest IT technologies can enable important advancements in the field. This can significantly lower the entry barrier for FAIR (Findability, Accessibility, Interoperability, and Reuse) data and ways of using or exploiting it within the energy community.

The gateway is directly connected to a data management tool, where datasets selected by EnerMaps can be viewed in an interactive way and analyzed for research purposes, or to support policy decisions, thus promoting the re-use of high-quality datasets in the field and create a bridge between data discovery and data analysis.

AI enhancing open data

Data management and the implementation of open science principles require a team of multi-skilled experts. For example, tools being developed by EnerMaps require both Artificial Intelligence (AI) experts and social scientists to combine the user experience, complex calculations and the generation of the energy data.

The importance of AI computing has been highlighted in the era of Big Data, with more and more researchers turning to machine-learning models to develop accurate, yet fast predicting models on a wide range of subjects. These models are widespread and constantly improving and increasing their possible applications. Yet one of the main challenges is still the process needed to transform data from its original raw form to one that is compatible with these learning models. For data scientists, much of the time is spent transforming these datasets, rather than on developing and improving the learning model itself. With this in mind, the EnerMaps team is using its interdisciplinary expertise to develop a platform that makes a wide range of datasets adapted to developments in AI.

New insights on energy data

Artificial Intelligence is often very computationally intensive during the training phase. However, the deployment of a final, fully-trained model is usually a much easier, less computationally intensive task. Model deployment is thus much more adapted to portable and lightweight applications, such as web platforms that are supposed to generate results almost in real-time.

Taking this into account, the EnerMaps web platform provides an environment for Calculation Modules (CMs) based on AI, facilitating the dissemination and deployment of machine learning models. The online platform will give access to the input datasets that are needed to generate predictions, for example of building energy consumption. In this way, the platform will allow users to have access to additional energy data, which is produced on the fly by AI-driven CMs, rather than by distributing pre-calculated data. The platform is therefore vital on two levels: for scientists, it allows for easy distribution of AI models and represents a significant step towards open and reproducible science; and for the end-user, it provides access to additional layers of energy information that are derived from the original datasets, thus enriching energy analyses with new insights that would not be possible with only the original datasets.

User-friendly data

In order to effectively meet the needs of energy researchers throughout Europe, there needs must be at heart of the development of the tools supporting data access. A case study conducted by the EnerMaps project is a perfect example of this focus on the human element, particularly during the dataset selection and quality-check (QC) processes.

During this study, stakeholders and those involved in the EnerMaps project took part in surveys and focus groups to determine the needs of energy researchers and professionals.

Using a sociological approach, the project conducted an initial survey utilized a stakeholder analysis to identify individuals whose experiences and needs helped guide the dataset selection process. The selected stakeholders were then asked to share their thoughts on energy data and to provide user stories that could be used to drive the development of EnerMaps in a direction that satisfied members across the energy research community.

Gathering human inputs was also done by interviewing experts in the field of energy in need of quality and relevant data. The interview phase is one of the basics in sociology. EnerMaps sociology researchers analyzed this feedback on their data needs and, based on these results, the researchers were able to determine some of the key areas where energy researchers demanded higher quality data.

A social network for the scientific energy community

One thing that we are slowly losing as a result of the internet is human contact. This is no-less prevalent in research, where having all the information online and open access means communication between contributors is being transformed. Researchers are losing contacts within their communities and are now looking to recreate a virtual community. One of the goals of the EnerMaps project is to create an active scientific community whose members can engage with one another, discuss data, their research, ask questions and help advance learning in research communities as a whole. To this end, the Kialo social network will be connected to the EnerMaps tool, giving users the space to facilitate this exchange of ideas and data, and help contribute to advancing our collective capabilities in advancing the energy transition.

Capacity building

While the creation of the EnerMaps is helping to make data more accessible, it is nothing without researchers using it. This makes capacity building programs vital – to inform, train and engage with future users.

The EnerMaps capacity building program has been developed to be extremely customizable, enabling developer to create a user-specific tool. Three main user target groups have been identified: research and academia, industry, and public administration, on top of additional users such as data providers and social innovators. A capacity building survey is now being used to consolidate such classification. This training approach aims at delivering to each user all the knowledge needed to maximize their benefit from using EnerMaps and specifically the functionalities relevant to their target group.

This capacity building process, together with human aspects highlighted in its creation of user-friendly data and a social network for researchers, is leading the way in showing what data-sharing can look like in the future – something that is tailored to ensure users can utilize and share data as quickly and effectively as possible.

Accelerating the energy transition

When asked to think about the energy transition, most people would imagine solar panels and wind turbines. While these of course make up part of the process, access to energy data is invaluable if we are to properly plan and implement future energy-saving projects that help Europe meet its climate targets.

Now, due to the increase exposure at the European level, ‘Open Science’ has grown more important and become more widespread in its use. But the more it grows, the more complicated it becomes, with increasingly large amounts of data available, but without the qualitative and user-friendly tools needed to utilize it effectively.

This is why European-funded projects like EnerMaps are so important. Already its expert-driven data selection process has provided a vast starting point with 50 different datasets. With this amount of data, and a platform for researchers to effectively access and share knowledge, progress in the energy transition can be made.