Data Inventories and Policy Landscape
This report presents a summary of existing experimental data holdings together with current access policies in European experiments. It also presents an assessment of their compliance with FAIR Principles and makes suggestions for how to improve the present practices/policies towards a more FAIR and more Open Data management. All European tokamak and stellarator experiments grant access to their measured and processed data on an individual basis, to collaborators who are formally identified as members of the experiment's team. Once a researcher is authorized for a given experiment, he has access to all measured data and processed data (Plasma Reconstruction Chain, PRC) of that experiment. Data has some degree of FAIRness at the level of a given experiment, but EU experiments are presently not interoperable, which prevents from exploiting results of the EU fusion experiments at their full potential. In particular, Data Mining / Machine Learning activities cannot be conducted across multiple experiments, or would require the ad-hoc creation of specific databases. A few international multi-machine databases have been created in the last decades of fusion research but their perimeter is limited to specific physics topics (e.g. confinement, disruptions, …) and they are not fed on an automated/systematic basis. In addition to improving the EU fusion science community Open Science and FAIR practices, making metadata and data interoperable across EU experiments is a key target of our recommendations since it would bring unique benefits to the EU fusion research, increasing the potential for new discoveries. The IMAS Data Dictionary is recommended as the standard ontology for achieving interoperability.