Modeling domain metadata beyond metadata standards

Abstract

The Laser Interferometer Gravitational-wave Observatory (LIGO) project to detect gravitational waves represents a complex, distributed scientific endeavor posing specific challenges for reproducibility and data management. The integration of provenance and other metadata information into the workflow stands as one means of addressing such challenges. The goal of a metadata model for the LIGO workflow is the provision of metadata describing all the data products at each significant milestone in the data analysis pipeline. Given both the highly specific domain and the need to support current analysis tools, the development of such a model demands a more complex, comprehensive approach. For this reason, we pursued a multipronged approach to metadata modeling, gathering users’ conceptions, system information, research artifacts, and other organizational documents, and worked to combine the findings into one final model. This approach provided a thorough understanding of the overall research lifecycle and insight into scientific workflow metadata modeling.