Jarves: The Digital Data Steward for Engineering Science Research

A presentation about Jarves on the NFDI4Ing Conference 2023.

Find Jarves on: https://jarves.nfdi4ing.de/

Abstract:
Rising amounts of data and the corresponding management and demands of funding institutions lead to an increasing significance of research data management (RDM), which introduces researchers to new competencies. Researchers must therefore know what specifications they have to meet with their RDM. There is also a demand for information on which tools are available to support them, and where they can receive guidance or training materials on RDM. However, a survey conducted amongst 168 researchers showed that more than half of the interviewees do not know the FAIR principles. Over 60% do not use a data management plan for their research. To foster RDM in engineering, researchers must be informed about the possibilities and practices of successful research data management.

However, researchers are often missing rules, guidance, or related knowledge for their RDM while facing additional effort because of missing automation and repetitive work. In the aforementioned survey, only 22% of the interviewees claimed to have access to RDM specifications, while 39% answered to not have any support at all. The latter statement shows a discrepancy between the interviewees’ perceptions and the existing solutions. For instance, several universities offer RDM support to help researchers get started on RDM. Also, there are training materials either by universities or initiatives such as the NFDI, respectively, included consortia. Conducted interviews showed that solutions and tools offered, as well as the proper usage of these, are often not known by researchers. Numerous researchers wished for more automation in research data management, as RDM itself is often perceived as an additional effort rather than as a benefit. The gap between RDM theory and research practice in the engineering community must be addressed in a way that combines guidance and applicability with existing solutions.

The high initial effort to define or research existing guidelines in RDM for researchers has to be considered along with the missing RDM knowledge amongst researchers. A new solution must be set to provide a benefit in RDM rather than more administrative work. As there are existing solutions, the goal should be to connect them rather than design a new one. To address these challenges, the status quo of RDM in engineering was surveyed, the needs of researchers were scouted via focus group interviews, and, RDM-process-related workshops were conducted. First, a survey on RDM in general and knowledge about it was performed among 168 researchers in the engineering sciences. The goal was to understand the spread and usage of RDM amongst engineering researchers. As a second step, focus group interviews were performed to identify subconsciously or consciously performed RDM tasks amongst the interviewees from engineering sciences.

For example, many manage their code via Git, which already is RDM but was often not considered as such by the interviewees. Thirdly, typical workflows and processes were collected in a workshop format. The workshops were designed to be as open as possible to not set false artificial boundaries to the researchers’ everyday work when designing an RDM process for them.

Lastly, a suitable solution was conceptualised, possible technical solutions were scouted, and a demonstrator was implemented. As a result, a framework, the “Joint Assistant for Research in Versatile Engineering Sciences”, or in short Jarves, was created that aims to address these challenges by providing a structured RDM process with recommendations regarding tools and training materials and partial automation. It is designed as a digital data steward, specifically for research data management in engineering. Jarves provides a guideline for RDM that can be tailored to the specific RDM rules of the researchers’ university, institute, or funding institution. The so structured process guides researchers through their everyday research data management, supporting them not only with steps and tasks to perform but also practical solutions and training materials for RDM issues, readily accessible at the point of need.

Jarves’ integrated decision support considers multiple factors for the hints it provides. For example, not only the institution’s RDM guideline influences the archiving solution. The specific disciplines involved, the amount of data collected, and the presence of personal data have to be taken into account. From the creation of the research data management plan to the archiving of data at the end of the project, Jarves helps researchers manage research data at every stage so that they can focus on their research tasks. Existing online tools in research data management, such as RDMO (Research Data Management Organizer) are integrated into Jarves via API access. This integration allows researchers to switch between different data management tools seamlessly while keeping them synchronized.

The proposed framework shall be demonstrated via the web tool Jarves to inform the engineering community on the one hand and enable the authors to collect feedback to further enhance the framework and its implementation in its development period. Therefore, the demonstration contains a brief overlook of the goal and purpose of Jarves, followed by technical aspects such as architecture and implementation approaches. The latter is meant to give other developers of tools an overview of how to make their tool ready to connect to Jarves, as well as how to connect to Jarves themselves. After that, the current status of the interface implementation is presented along with certain functionalities, including the steps of the engineering workflows, the concept of guidelines and overwriting, the decision support, and the connection to other tools. The feedback during the proposed demonstration will be used to enhance Jarves’ underlying framework and its implementation regarding its effectiveness in guiding researchers through their research data management. Further tools will be integrated after initial feedback to provide a widely available and interoperable platform in the field of RDM in engineering. An extended evaluation will take place at a later time to validate the framework.