Flashtalk for Open Science Seminar, June 2025

presentation

A 3 minute flashtalk introducing our work in the Seedcase Project. Presented at the Open Science in Health Research: From Principles to Practice Seminar, June 2025.

Author

Signe Kirk Brødbæk

Published

June 26, 2025

Note

This is the transcript of the flashtalk presented at the Open Science in Health Research: From Principles to Practice Seminar, in June 2025.

The brief:

  • You will have 3 minutes to present your project/resource
  • Highlight how open science principles shaped this work

Hi, I’m Signe, a software engineer at Aarhus University and Steno Diabetes Center Aarhus. I’m here to tell you about a project I work on called Seedcase.

So, what is Seedcase?

Seedcase is a data infrastructure initiative focused on building open, scalable tools for working with research data. From the start, our work has been grounded in open science principles. Everything we build is open source and developed in the open on GitHub under our GitHub organisation. We also rely on other open source tools for our work, such as the Parquet data format and the Frictionless Data Package standard. Lastly, we actively document not just our software, but our processes, values, and design decisions, following a documentation-driven development approach. This documentation, both internally and externally facing is published openly as websites for anyone to explore, reuse, and contribute to.

What are we actually building?

The core of Seedcase is a set of Python packages. We want to simplify the process of going from messy raw data to structured and documented Data Packages and make it findable via a website. Data Package is a format we’ve adopted from the Frictionless Data Package Standard that’s designed to make data FAIR: Findable, Accessible, Interoperable, and Reusable.

So, our first package, Sprout, which will soon be published to PyPI, helps you create Data Packages by simplifying the process of structuring and checking your tabular data and metadata according to the Data Package standard.

We’ve also planned:

  • A metadata checker, check-datapackage, which ensures your data and metadata are consistent and complete. We plan to release this as a standalone package to support any project using the Data Package standard, to promote reuse beyond our own tooling.

  • A template project, template-data-package, to help new users getting started with building their Data Package. This includes the recommended structure and dependencies of Data Packages following the approach built by Sprout.

  • Now that we have a structured Data Package, we also want to make it findable. For this, we’ll create a metadata website generator, Flower, which lets researchers turn metadata into a simple, shareable website. This makes it easier for others to discover and understand your dataset, without the need to download anything.

What’s next?

This fall, we’ll be collaborating with two large research projects at Steno Diabetes Center Aarhus to test and apply our tools in real-world settings. These partnerships will be a key milestone by seeing our open infrastructure used to help researchers structure their data more transparently.

Lastly, while I’m presenting today, Seedcase is a team effort. We’re two software engineers, a database administrator, and a team lead—all are committed to building open tools for open science.

Thanks for listening! And feel free to check us out or contribute on GitHub!

Slides