Contents of data management plans and RDM proposal chapters

The generation of new data or the reuse of existing data in your projects should be documented as soon as it occurs. This way it is always clear from where the data originates from or how it was created. Each subsequent processing step should also be recorded. It is advisable to use structured database recording systems instead of pure continuous text descriptions in order to be able to search efficiently and reliably for specific information in this documentation at a later date.

There are standards for recording such metadata that define the exact format for certain information (e.g. authors, survey date, project context, etc.). The naming of other fields and categories that are not predefined in such a metadata schema should be based on subject-specific standard vocabularies (terminologies and ontologies), if these exist. This way, you increase comparability and interoperability with data of a similar nature from other projects.

There are many potential sources of error when collecting and processing data, such as incorrectly calibrated measuring devices, transmission errors, corrupt files and all conceivable forms of human error. As incorrect data also leads to incorrect research results, regular quality checking mechanisms are important. Depending on the type of data, these measures can sometimes be (partially) automated. For example, creating and comparing checksums helps to recognize files that have been damaged or unintentionally changed during copying. Programmes such as Excel are known to change character encodings and number formats without prompting, which is why a comparison between the original and the Excel version is recommended when converting. If possible, transcriptions and annotations should always be checked independently by at least two persons.

In principle, scientific data should be made available in such a way that it can be read with standard software. Proprietary data formats that can only be read with special and possibly cost-intensive software should therefore additionally be converted into open formats if possible, even if it sometimes results in a reduced functionality. At the very least, however, it should always be specified exactly which version of a programme and, if applicable, also which operating system is required.

The interviews are processed with the free software EXMARaLDA according to the standards of the ‘Gesprächsanalytisches Transkriptionssystem 2’ (GAT2) and provided with a corresponding header. After the automatic, AI-supported (OpenAI Whisper) transcription of the interview recordings, a quality check is carried out by one of the project participants. The audio recording is played back, and the transcript is read at the same time. Incorrectly transcribed content is corrected manually. In addition, pauses and non-speech sounds such as laughter, coughing, clearing of the throat, etc. are recognized. The finished transcripts are additionally converted into the RTF format, which can be read with any standard word processing programme.

All components are designed digitally with SOLIDWORKS. The industrial partner provides an instance of SOLIDWORKS Product Data Management (PDM), in which all data is versioned and its processing history documented. SOLIDWORKS Simulation is used to simulate the mechanical stability of each component to identify incorrect calculations before a physical prototype is produced. At the end of the project, the documentation for each component is exported as a static PDF file and the CAD file itself is exported into the STEP format (ISO 10303).

Vector data is of the shp format, raster data of the Geotiff or csv formats. A PostGIS geodatabase is created additionally. These formats can be integrated into any standard geographic information system. Vector data is checked for obvious geometric inconsistencies using the QGIS plug-in ‘Geometry Checker’ and corrected if necessary. All metadata is recorded in accordance with ISO standard 19115. Methods for processing and analysing the geodata are described in the project wiki (Doku-Wiki), which is included in the scope of the LUIS Projektablage.

When handling research data, often affected areas of law are data protection and personal rights as well as copyright and patent law. In order to avoid any violations of law, check whether and which legal standards or existing contractual provisions may be affected before reusing, collecting, processing, evaluating and, if applicable, archiving and publishing data. Pay particular attention to license conditions, declarations of consent, non-disclosure agreements, contracts for commissioned processing and cooperation agreements. If necessary, make contractual agreements yourself. When reviewing existing legal texts as well as when drawing up new contracts, you should always consult legal experts!

In principle, most major funding organizations now expect data from the projects they fund to be made publicly available for subsequent use, provided there are no legal or ethical reasons to the contrary. However, access to data must always be restricted or even completely denied if third-party rights would otherwise be violated. This is regularly the case with personal data, for example, unless the data subjects have given their legally valid consent to publication. Non-disclosure agreements with (raw) data providers, certain licence conditions, commercial interests or ethical considerations can also rule out publication or restrict access. If you generate new data collaboratively, all parties involved should agree to the planned publication. It is therefore advisable to reach written agreements before starting a project.

In this context, it is particularly important to consistently document the sources or the generation of data as well as licence conditions and rights holders, if applicable. If you want to process data from external sources for which no precise licence conditions or other re-use agreements are known, obtain all necessary usage rights from the rights holders in writing before using the data. Reach contractual agreements within your project team on the extent to which data generated in the project may be used by all participants. Also take into account the possibility of participants leaving the project prematurely or changing employers. If you publish data yourself, attach a licence with unambiguous and legally binding conditions for re-use.

A large part of the ‘Guidelines for Safeguarding Good Research Practice’ refers to the handling of research data. More specific requirements may be laid down in discipline-specific guidelines and codes, but these are not always widely known. In any case, check whether documents relevant to your subject exist whose contents should be taken into account. The DFG has compiled subject-specific recommendations on the handling of research data, most of which have been published by the review boards or major research associations. On forschungsdaten.info there is a subpage on RDM guidelines, including discipline-specific guidelines. Another compilation can be found at forschungsdaten.org (German only). Many NFDI consortia now offer helpdesks where you can also enquire about current standards relevant to your discipline.

The project compiles and analyses excerpts from published texts as well as unpublished archive material. A Zotero literature database will be used to systematically record the sources, copyright holders and usage conditions for all texts. An expert legal examination concluded that the use of the published texts in the form envisaged in this project is covered by the copyright limitation provisions. In contrast, the conditions for the use of the unpublished material vary from archive to archive and other rights holders, in particular the authors or their heirs. Attempts will be made to reach contractual agreements with the respective parties that not only permit the evaluation but also the publication of these texts, possibly in a modified form (e.g. anonymised or paraphrased).

Experiments on mice are indispensable for the realization of the project. A corresponding authorization will be applied for from the Lower Saxony State Office for Consumer Protection and Food Safety. For ethical reasons, we limit the number of experiments and the animals used in them to the absolutely necessary minimum in accordance with the 3R principles. The experiments are designed in such a way that the animals are exposed to as little stress and pain as possible. The total of 20 animals required from the patented breeding line C57BL/6J are used and kept in accordance with the licence conditions of TACONIC BIOSCIENCES.

In accordance with the funding conditions for this tender, the construction plans for the innovative wind turbine developed in the project are published under open licences such as MIT or CC0 and can therefore also be used commercially free of charge. However, some of the installed components are provided by external manufacturers and subject to their licence terms. The exact circuit diagrams for these as yet unpatented parts were made available to the project by the manufacturers under the condition of confidentiality and non-disclosure. The plans and technical descriptions intended for publication therefore only represent these components as a whole and without details of the internal structure. Furthermore, these documents will be submitted to the external manufacturers for review prior to publication.

Ultimately, the respective management of a research project is responsible for the appropriate handling of research data. However, they can and should delegate tasks and responsibilities, for example to principal investigators, doctoral students or student assistants. Depending on the size of the project and the complexity of the data management, it may make sense to employ a person who takes on the role of a full-time data steward and dealing with the administration and documentation of the data, as well as for setting up and operating the necessary systems. The operators of storage infrastructures should guarantee their secure operation. External service providers should be obliged to maintain confidentiality and comply with security standards by means of a commissioned processing contract.

Data management mainly requires resources in terms of personnel, infrastructure and external services. Ander all circumstances, the corresponding costs should be realistically estimated from the outset, budgeted and, if necessary, applied for. A full-time data steward could, for example, correspond to a TVL-13 position. The calculation becomes more difficult if the data management is to be carried out entirely by the scientific staff, who then have to reserve part of their working time for this task. On the infrastructure side, the required storage space in the short and long term should be considered. In this respect, costs may not only increase due to volume, but also because of special requirements, for example in terms of security. Adequate infrastructure is not always sufficiently available within the university's basic facilities. External services can sometimes be used already at the stage of data acquisition and processing. However, costs are most frequently incurred for long-term archiving, curation and, if necessary, public provision of data in external archives and repositories.

At the simplest level, curation means simple bitstream preservation, i.e. the guarantee that files remain unchanged. If the data is stored in a professional repository or data archive, this aspect falls under the responsibility of the relevant infrastructure operators. The project director should ensure that all relevant data is stored in such facilities at the end of the project. Active curation, on the other hand, may also include quality checks, enrichment and checking of metadata and regular conversion to newer file formats. Some repositories and archives offer such services (for a fee). Curation data long-term on your own is not advisable.

The responsibilities for handling research data are set out in more detail in an internal guideline of the joint project. From among the PIs, Prof X is appointed the primary contact person on the topic for the funding body and external enquiries. We apply for a TV-L 13 full-time position for a data manager who will configure and maintain the systems for data and knowledge management jointly used by all partners and who will be available to all members of the network as a contact person for practical questions. The computer centre of University Y provides the technical infrastructure, including the backup servers, and is responsible for its technical functionality and security. The responsibility for the correct processing, documentation and storage of data at the different partner institutions lies with the respective sub-project managers and, regarding the technical operation, with the local infrastructure operators.

The applicant is responsible for the data management in general. She will instruct the doctoral researchers in her junior research group and will randomly check during the course of the project whether they have named, filed and documented their data in accordance with the data management plan. Due to the expected data volume of approx. 5 TB, a fee must be paid for the use of the High-Seas cloud service, which is operated by LUIS. The fee will average around 500 euros per year and will therefore amount to around 2,000 euros over the four-year project duration. At the end of the project, the data will be published in the open and long-term readable file formats csv, tiff and png in the established disciplinary repository Pangaea, which is operated by the Alfred Wegener Institute (AWI). The AWI is therefore responsible for the long-term preservation and availability of the data. No curation of the data is required beyond a mere bitstream preservation.

The applicant is responsible for the correct handling of the research data processed in the project in line with disciplinary, legal und technical standards. He carries out all processing and documentation steps himself and allocates approximately 10% of his weekly working time for this task. If needed, he may seek professional advice from the LUH central Team Research Data Management or the helpdesk of the NFDI consortium X. The IT administrator of the Institute Y is responsible for the technical operation and security of the institute's server used as storage location. The operation and security of the backup servers are ensured by LUIS. At the end of the project, a final quality check, long-term curation, and publication of the data will be conducted via the paid GESIS data service “Premium Archiving”. A corresponding tender amounting to XXX euros has been obtained.