
Developing code that can be reused and shared amongst collaborators is a growing part of an astronomer’s job. In 2024-25, 36% of the published papers in Physics and Astronomy mentioned software or code in their text. Many journals now require or at least encourage that the code underlying scientific results be made public and open-source. For astronomers to publish high-impact research, good coding practices for creating openly available software are essential. This requires advanced statistical methods and their implementation into software. ASTRODAT will help late-stage PhD students and ECRs in Astronomy come together and develop these skills.
Do you often feel like you had some training during your PhD, but now that you're past the first few years, you need the next level of training? This workshop is for you! We are aiming to build on the foundation provided by the workshops aimed at first-year PhD students.
The ASTRODAT workshop will help you make the most of the huge amounts of data from facilities such as ESA Euclid, Rubin LSST, DESI, ELT HARMONI, and the SKA, which will drive unprecedented progress in the next decade. Space telescopes like JWST, Hubble, Kepler and Galex also contribute to this abundance of data. On top of the petabytes of data produced by observatories, cosmological and astrophysical simulations already produce similar amounts of data while requiring 107 CPU hours to run. The processing and modelling of this wealth of data require advanced programming and data science skills, such as machine learning/AI. ASTRODAT will offer specific training in these areas.
The week-long workshop will contain two components:
1. Hands-on lectures and workshops each morning, building skills in statistics and code development with topics such as bash, slurm, version management through git, Python package deployment, Bayesian inference, parameter sampling, interpretable machine learning methods (e.g. simulation-based inference), etc.
2. Collaborative hack sessions each afternoon, putting into practice what was learnt in the morning with two choices:
- Developing a new Python package over the week in a group
- Contributing to existing open-source code - Astropy
Earlybird Registration: TBD
Registration Ends: TBD
Workshop Dates: 8 - 12 September 2025
COMING SOON!