For this 5th edition, the pySTEPS community joins forces with the open radar and MLCast communities in the Open Nowcasting Hackathon 2026. This hackathon is set for Mon 2 to Fri 6 November 2026 at "Chez Philippe" near the three-country-point (BE/DE/NL). Apply here by 16 October: https://forms.gle/g8x3VMTSAtux3oSB6 

One topic we’d especially like to explore is how MLCast, pySTEPS, and open radar tools such as xarray can work better together: for example, making it easier to use MLCast models and workflows within the pySTEPS ecosystem, exchange data and outputs between them and xarray, and identify useful interfaces or common standards.
More broadly, the idea is to spend a few days together sharing ideas, coding, discussing research, and hopefully creating new collaborations across the different nowcasting communities.

The programme will be shaped together with everyone taking part, including:

• Improving interoperability between MLCast, pySTEPS and xarray
• Implementing machine-learning and deep-learning approaches to nowcasting
• Creating AI-ready radar, satellite, and other observational data
• Prototyping new features and research ideas
• Closing open issues and pull requests
• Improving interoperability between other open-source nowcasting tools
• Documentation, examples, and tutorials
• Lightning talks, technical discussions, and plenty of collaborative coding

The goal isn’t just to write code. We’d also like this to be an informal space to exchange ideas, learn from each other, and explore opportunities to work together.

Important to know: This hackathon is not a basic training, familiarity with (some of) the software packages is required, so please come prepared.

We are grateful to BELSPO for their financial support through the DERISC project (Royal Meteorological Institute of Belgium (RMI) - Vrije Universiteit Brussel - KU Leuven). This unique format wouldn't be possible without it.