Leibniz Data Manager : an Adaptive Research Data Management

Current scientific work is reflected, among other things, in an increasing number of publications that hardly anyone can keep track of [1]. In addition, there is a growing demand for the reproducibility of scientific findings, not least since the so-called replication crisis. In the course of this, more and more research data from which the new findings were derived is published. It is becoming increasingly difficult for researchers and others involved in research to maintain an overview, even in their own field of research. Within the project LDM-Explore we offer the Leibniz Data Manager (LDM), a tool to search and explore research data across different repositories and evaluate their potential for re-use [2]. Since 2017, a prototype is available in cooperation with Leibniz University Hannover, which is being continuously expanded since September 2020 in our project. LDM can be found and used at https://service.tib.eu/ldmservice/. With LDM we developed an adaptive research data management system, based on the open web-based data catalog software CKAN (Comprehensive Knowledge Archive Network) . LDM offers a simple and user-friendly interface which is combined with many optional features, e.g. harvesting, data visualization and preview, full text and fuzzy search, and faceting. This can help researchers to better find and evaluate data from integrated repositories. Since LDM is available as a Docker  container, customized local LDM distributions can be installed. The functions of LDM are closely related to the FAIR principles. The heterogeneous research data in LDM is findable via a PID system, here DataCite DOIs, which leads to a landing page in LDM. Authors can specify their ORCIDs and email addresses. The data is described by metadata in several schemata and vocabularies: DCAT, DataCite, and DublinCore. The data type is automatically recognized and specified. The datasets can be explored by keyword queries or searches based on DCAT properties. The accessibility of (meta)data in LDM is reflected in open access in principle, but this can be restricted for the data if desired, which is important especially for sensitive data. Data can be explored by generating a preview of the data in LDM. This way, researchers can decide whether the dataset is relevant for them or not before the download. LDM supports the visualization of AutoCAD files, different views on the same dataset, e.g. in 2D and 3D, and the execution of live code via Jupyter notebooks. The interoperable data in LDM is described by metadata e.g. in DCAT in several RDF serializations. This metadata can be exported and integrated in other systems. Data from other repositories can also be automatically integrated in LDM. Different synchronization schedules can be set to keep datasets up-to-date. Finally, the reusability of (meta)data in LDM is achieved through open licenses (CC0 for metadata), versioning, citation suggestions (export in CSL, BibTeX), and by the possibility to download or export both the data and the metadata. The shown FAIRness of the stored (meta)data in LDM also impacts the Research Data Management, which affects all stages of the research data lifecycle. This is reflected in particular in the aspects Publishing and Sharing as well as Reusing Data. We strive to continuously increase and improve the FAIRness of data in LDM. At the poster session at the workshop we want to demonstrate the crucial role of metadata for the effective and efficient analysis and management of research data.