Choosing a Repository
Choosing a repository depends on many factors, such as whether your files should stay private, if it is ongoing work, or if you have specific needs. As OSF is changing, we try to give some guidance on the current repository landscape. Contact us for help with your specific situation.
Research data repositories can be general-purpose repositories (for example Zenodo, Dryad, Figshare and Dataverse installations), disciplinary repositories (see language archives listed below), institutional repositories run by universities and libraries, and also e.g. repositories run by journals. It is generally a good idea to use a disciplinary or alternatively an institutional repository. Zenodo, run by CERN, is possibly the most widely used general repository.
There is already very sophisticated infrastructure available for language research data. CLARIN, the Common Language Resources and Technology Infrastructure, is an institution that consists of a network of centres that provide repositories and services for language resources. Check the relevant centre profile before choosing one. Germany's NFDI consortium Text+ is very similar to CLARIN and incorporates many German CLARIN centres. In short, CLARIN is the European network of centres, and Text+ is the German consortium that coordinates language and text data services, including many of those same centres. A CLARIN Centre is an institution, such as a university data centre or a speech archive, that offers repositories, tools, or expertise for language resources and is part of the CLARIN network. There is no single CLARIN upload site: each centre decides which data it accepts and how deposit works. The centre registry lists every centre with its type, and other details. Some are knowledge centres focused on a topic such as learner corpora or treebanks.
Distinguishing Different Goals
- Ongoing work: Use your SFB project 2TB project box on Sciebo for shared online storage. Do not use it for sensitive data (Datenschutz-Grundverordnung).
- (Large) research data: The newly established DataStorage.nrw provides 100 GB of storage automatically and more on request, see the ITCC service page. This should be ideal for NRW based researchers. The ITCC storage and sharing overview also lists other options such as SOFS and backup tools.
- Transferring (large) files: GigaMove 2.0 lets you easily share data up to 1 TB. The service is available to most European researchers, see: DFN-AAI entity list.
- Code: University of Cologne GitLab supports version control and collaboration, with most features that are also available on GitHub. Non-textual data is better stored in a repository or shared folder.
- Sharing finished work: Ideally choose a disciplinary-specific repository. See the options below. Think carefully about the access and usage rights of your data.
- Archiving finished work: All projects in the SFB use the pipeline provided by the Data Center for the Humanities.
Language Research Data Repositories
- Language Archive Cologne (CLARIN ? Centre) is used mainly for language documentation, recordings, and annotations. It is managed by the Data Center for the Humanities (DCH), who also advise on preserving and presenting humanities data, including multimodal material.
- BAS (CLARIN B Centre), hosted by LMU Munich, focuses on spoken-language data.
- IDS Mannheim's Database for Spoken German (DGD) provides access to spoken German corpora.
- TROLLing is a repository for (applied) language and linguistics research data.
- IRIS