Besides my scientific projects I am also involved in the preparation and presentation of tutorials showcasing how to best find, discover and access publicly available data. I furthermore give input about how I as an astronomer think archives of radio astronomy data and future science platforms should look like.

Software and Data Science skills


  • I have started to contribute to the Astropy package with tutorials
  • I regularly use astropy, astropy affiliated packages, numpy, scipy, matplotlib, seaborn, pandas, scikit-learn and keras on tensorflow

Data Science Projects
I have started some small personal data science projects one of which is a look at the CoViD-19 cases in Germany. I have also started to dabble with forecasting the weather in some German cities. Currently it seems my machine learning algorithm refuses to learn about climate change. Once I convince my algorithms that climate change is a thing, you'll find the results here.

Other Software Packages and Language I use

  • Aladin, TOPCAT, ds9
  • Miriad, IRAF, GILDAS
  • GIMP, Inkscape

As the Asterics post-doc my main task was to co-organise the fourth Asterics VO school (one of the deliverables of this Horizon 2020 project) together with my colleague Ada Nebot. Even after the Asterics project has finished in May 2019, I continue to work on tutorials that showcase the CDS services and the VO in general. Our current focus lies on Python based tutorials as more and more astronomers rely on Python for their research. You may find these tutorials here and here. More VO tutorials that use other software packages than Python can be found on the EURO-VO page

While creating tutorials in Jupyter notebooks and during my day to day scientific work, I have many chances to test and play around with Jupyter notebooks. From this experience I can inform on how future science platforms should look like.

With the rise of new telescopes such as MeerKAT, ASKAP, SKA, LSST ect. the amount of data that is available to every astronomer increases to an unprecedented volume. These data can not be analysed on a personal computer any more. This is where science platforms come into play: these platforms follow the principle of "code to data", where scientist run the computationally heavy parts of their analysis pipeline on remote machines.

So far I find that many day to day tasks (also computationally heavy tasks) that astronomers need to tackle are already implemented in well maintained Python packages. These could immediately be used on science platforms, which are likely based on Jupyter notebooks with Python kernels. There are, however, some tasks like for example galaxy SED modelling that are not (yet) available as Python packages. For these software packages the community needs to decide, whether these packages needs for example Python wrappers, so that they could be used on science platforms or whether these packages can remain running on local machines in the future.

So far the radio astronomy community does not rely as strongly as other communities on the Virtual Observatory and its protocols. The VO protocols help to unify finding and accessing data across different archives. I am currently involved in many discussions on how VO standards and protocols may best be implemented for radio data archives and how I as an active astronomer would like these services to be. To this end I have also joined the recently formed Radio Astronomy Interest Group in the IVOA.