Search

FAIR handling of scientific data plays a significant role in current efforts towards a more sustainable research culture and serves as a prerequisite for the fourth scientific paradigm, that is, data-driven research. To enforce the FAIR principles by ensuring the reproducibility of scientific data and tracking their provenance comprehensibly, the FAIR modelling of research processes in form of automatable workflows is necessary. By providing reusable procedures containing expert knowledge, such workflows contribute decisively to the quality and the acceleration of scientific research. In this work, the requirements for a system to be capable of modelling FAIR workflows are defined and a generic concept for modelling research processes as workflows is developed. For this, research processes are iteratively divided into impartible subprocesses at different detail levels using the input-process-output model. The concrete software implementation of the identified, universally applicable concept is finally presented in form of the workflow editor KadiStudio of the Karlsruhe Data Infrastructure for Materials Science (Kadi4Mat).

Scientific processes produce huge amounts of data that are usually acquired, transformed and analyzed on a regular basis. Translating these processes into automatable and reproducible workflows is considered to be an efficient way to support scientists in performing repeated processes that would otherwise be time-consuming and error-prone tasks. Consequently, the quality of scientific research can be accelerated and enhanced. In this article, we present for the first time a use-case of KadiStudio as a tool to automate analysis procedures of scientific data that are repeatedly acquired from in situ scanning electron microscope (SEM) micromechanical testing. KadiStudio provides a desktop-based workflow editor as part of the ecosystem of Kadi4Mat: Karlsruhe Data Infrastructure for material Science. The presented workflow includes nodes for processing and analysis of different types of data, namely mechanical response in text format and a series of SEM images in video file format acquired during in situ SEM deformation tests. In addition, the raw and analyzed data are automatically uploaded to the KadiWeb repository via nodes based on the kadi-apy library.

The concepts and current developments of a research data infrastructure for materials science are presented, extending and combining the features of an electronic lab notebook and a repository. The objective of this infrastructure is to incorporate the possibility of structured data storage and data exchange with documented and reproducible data analysis and visualization, which finally leads to the publication of the data. This way, researchers can be supported throughout the entire research process. The software is being developed as a web-based and desktop-based system, offering both a graphical user interface and a programmatic interface. The focus of the development is on the integration of technologies and systems based on both established as well as new concepts. Due to the heterogeneous nature of materials science data, the current features are kept mostly generic, and the structuring of the data is largely left to the users. As a result, an extension of the research data infrastructure to other disciplines is possible in the future. The source code of the project is publicly available under a permissive Apache 2.0 license.