Improved Discovery and Access to Research Data in Energy Systems Analysis

Kurzfassung

Introduction Research in the domain of energy systems analysis deals with the evaluation of future sustainable energy systems. This is often done by computational modelling of the energy flows, trading at energy markets and investment decisions of the various actors within the energy system. The analysis is therefore driven by data to very large extend. The modelling uses a lot of different input datasets for large variety of domains as e.g. engineering, economy, climate and societal developments. The models itself produce a lot of data as a result which contains energy balances, time series of energy provision or trading on energy markets, possible future investment needs into energy technology and much more. To be able to model the complex interactions of energy, markets and society, models more often get chained, so the output of one model becomes the input to the next model. Therefore, a lot of data is used and produced within the research domain by very heterogenous actors. Improving findability Within the domain there is move towards more open data, but even then, all the datasets are stored on different places, in different formats with better or worse descriptive metadata. Within our project LOD-GEOSS, we try to use some lessons learned in the development of the Global Earth Observation Systems of Systems (GEOSS) and from linked open data (LOD) to improve the findability of all the research data. A central element is the DBpedia Databus which serves as a central communication platform and searchable catalog for the research data in our domain. If research data is made available suitable metadata along with a link to the dataset is registered to the Databus. The registration generates a unique id for the dataset which can be used to identify a specific dataset in the future. The metadata description and link to the data file improves the findability of the produced research data. A user of this data file can access the data though the provided link and reference this dataset by the unique id provided from the data bus. If a user republishes improved or derived data, a link to the originating data source is created. Data processing therefore gets traceable back to the originating source data. This can be complemented by data about the involved agents and activities to form a provenance graph of the data. The Databus and the distributed data storages therefore form a network of federated data bases of research data within our domain and improve the findability and access to the research data which is produced and used in our research area. Improving interpretability As described above, the data comes from a variety of different domains and interpreting the data is often a difficult task. In a parallel stream we contribute to the development of the Open Energy Ontology for the annotation of research data. This will create a common understanding of different data fields within the datasets and will ease data exchange within and across domains. Additionally, if data on the Databus is annotated with the ontology it will enable semantic searches for research data. With our presentation we want to show the current status of the development of a distributed data architecture to improve findability and access to our research data. This research is funded by a grant for German Ministry for Economics and Energy by grant number 03EI1005A.