André Langer, Lukas Schmolke, Martin Gaedke
BIRD: Using Conversational User Interfaces to Provide Relevant Metadata for Interdisciplinary Research Data Publishing
By the digitization of science, publishing research data to the scientific community is increasingly demanded in order to allow the replay, reuse or repurpose of existing research results. It is often necessary to describe the original research data files to increase its findability according the the FAIR principles for good scientific practice. So far, this is typically done by providing an additional descriptive floating text or by stating specific information in a submission form of a research data repository. Alternatively, structured machine-readable metadata description files can directly be provided. However, the meta description result quality depends on the experience of the user, commonly focuses on general aspects, and tool assistance is often limited. To address these issues, we applied as an alternative a conversational user interface paradigm to the description of research data and present the BIRD prototype (Bot-based Interface for Research Descriptions). It realizes a chat dialog which will assist scientists in providing an appropriate, structured metadata description based on the OpenAIRE Guidelines for Data Archives, independent of a particular technical repository platform. The tool is offered as a demonstrator for public access.