The DataHub Workshops
We regularly host workshops for all DataHub users. To give you an idea of what we offer, a list of our recurring series is included below. For our upcoming schedule, please see the next section. If the topics you are interested in aren’t currently available, we highly recommend browsing our past workshop recordings.
DataHub Crash Course: Services and Workflow
This course will provide an overview of our common data management workflow, our technical platform and our services. It conveys the minimum knowledge to start using the DataHub for versioned storing, sharing and publishing of your research code and data. It will also provide an overview on high performance compute resources and Jupyter interactive compute.
Handling Sensitive Data
This course covers the fundamental concepts of managing sensitive data, with a specific focus on our infrastructure: TAM GitLab and the High Performance Computing Cluster (MaRC3). Prior experience with Git version control or completion of the DataHub Crash Course is highly recommended.
Important: Completing this course is mandatory for anyone planning to work with sensitive data on DataHub platforms.
Git, Git-LFS, GitLab Basic Workshop
This course will enable you to use version control with git in a local and distributed/collaborative context via our TAM GitLab. The course will provide a basic understanding of git and its most important commands. It will include concepts of branching and merging parallel streams of work and will enable you to judge on advantages and limitations in using git version control.
Using Git-LFS you will learn how to manage even large amounts of research data efficiently. Demonstrating the usage of GitLab groups, visibility and memberships, we will enable you to collaborate in multiple distributed teams and projects.
Open Science, RDMPs, Metadata and Publication
This course aims at making your research process as smooth as possible while making your research products as visible and valuable as possible. First, we will look at the most underestimated phase of research data life cycle: Planning data management. We will cover a broad understanding of metadata, the interoperable and sustainable choice of data formats and platforms and finally the definition of re-usable pieces of work. We will discuss how this is reflected in research data managing plans (RDMP) and demonstrate how to set up an RDMP using the RDMO software.
These early investments pay off when it comes to publishing your research. We will discuss advantages and limitations of typical publication platforms like OSF, Zenodo, institutional repositories and subject-specific repositories like our own DataHub Repository. You will learn how to bundle and publish meaningful, richly described and independently reusable research objects (data, code, textual publications) with persistent identifiers (DOI). To increase findability and re-usability, we will cover typical licenses and how to interlink your work with other published work using defined relations.
High Performance Computing Cluster (MaRC3) and Jupyter Workshop
In this course we will practice how to scale up your computation using the MaRC3 high performance compute cluster. You will learn how to use git/GitLab from this central compute environment, manage software and dependencies using virtual environments and how to use the Slurm scheduling system to submit real-world compute jobs. We will have experienced HPC users who give first-hand advice on using compute resources effectively and avoiding typical pitfalls. We will discuss showcases to demonstrate the capabilities of high performance compute.
With our Jupyter_hpc service we also have a modern solution for browser-based interactive compute with various programming languages. As the Jupyter_hpc actually runs within the MaRC3, and includes the available filesystems and software, it bridges the gap between local and central code execution. You will learn about the Jupyter notebook format as a very powerful way of simultaneously generating, visualizing and describing research data in interleaved sections of code/results and text/documentation. We will demonstrate how to use the considerable capabilities of Jupyter_hpc for your research and discuss in which cases conventional HPC usage is more advisable.