Towards Porting Astrophysics Visual Analytics Services in the European Open Science Cloud

Abstract

The European Open Science Cloud (EOSC) aims to create a federated environment for hosting and processing research data to support science in all disciplines without geographical boundaries, such that data, software, methods and publications can be shared as part of an Open Science community of practice. This work presents the ongoing activities related to the implementation of visual analytics services, integrated into EOSC, towards addressing the diverse astrophysics user communities needs. These services rely on visualisation to manage the data life cycle process under FAIR principles, integrating data processing for imaging and multidimensional map creation and mosaicing, and applying machine learning techniques for detection of structures in large scale multidimensional maps.

Cloud-agnostic architectures for machine learning based on Apache Spark

“…The descriptors can further be configured and extended according to the needs of other scientific communities and/or commercial companies. The results of this work are currently utilized in several EU funded H2020 research projects (COLA [13], CloudiFacturing [32], NEANIAS [33]) as well as in national initiatives (ELKH Cloud and Artificial Intelligence National Laboratory [8]).…”