It is our pleasure to announce the fifth annual ESAC Data Analysis and Statistics (EDAS) workshop that will be held at the European Space Astronomy Centre near Madrid in Spain from Monday 2nd December to Thursday 5th December, 2019.


The primary aim of the EDAS workshop is knowledge transfer: to teach and instruct through lectures and guided hands-on exercises essential notions in data analysis and statistics, as well as modern techniques and methods to improve the way we treat data and do science.


This year the main focus will be on Machine Learning for astronomy and astrophysics, and our invited tutor is Siraj Raval.


The number of participants cannot exceed 150, and therefore we may have to carry out a selection process. If you are very interested, please register as early as possible.


We strive to have the most fruitful and useful workshops, and therefore encourage the greatest balance in diversity. In particular, we strongly encourage mid-career and women scientists such that we can strive to have as many men as women, and as many young as more seasoned scientists.



Registration opens on Friday 7th June 2019 and closes on Friday 4th October 2019. There will be a maximum of 150 participants. Depending on the number of registrations, a selection might have to be apllied.

The registration fee is 100 euros. This is non-refundable except in the case of extreme circumstances. Unfortunately no financial support is available this year. Lectures will be live streamed and later made available on the EDAS YouTube channel.

People working at ESA and CAB do not have to register.



Siraj Raval

Siraj Raval is an AI educator, bestselling author, and data scientist. He is founder of the international nonprofit School of AI based in over 400 cities globally, has built the fastest growing AI community in the world, and has worked on a host of open source work. Besides being a programmer, Siraj is also a speaker, rapper, and postmodernist.





Dr Guillaume Belanger 

Guillaume Belanger did his PhD on the study of the characteristics of the high energy emission in the Galactic Centre, and from Sgr A* in particular, working with XMM and INTEGRAL data. Since then, he has worked on developing statistically motivated methods for detecting transients in space, time, energy and frequency; more sensitive periodogram statistics for detecting weak low frequency periodic signals, and studying the shape of non-sinusoidal pulse profiles; and techniques for modelling red noise and characterising the properties of time-domain signals, all with a particular interest in variable X-ray and gamma-ray sources. Dr Belanger is the coordinator of the INTEGRAL Science Operations Centre at the European Space Agency.



The topics that will be covered this year include:

  • Machine learning basics, feed forward neural networks, optimisation, metrics
  • Regression & clustering algorithms (k-means, gaussian mixture models)
  • Classification (gaussian processes, SVMS, decision trees, neural networks)
  • Parameter estimation (hidden markov models, graphical models)
  • Mixture density networks, principle component analysis, generative adversarial netwroks, auto-encoders

Sessions will consist of both lectures and hands-on tutorials, including how the methods discussed can be generalised to other contexts.


For the tutorials, the following datasets will be used:

Gaia - Gaia is an esa mission designed to chart a 3D map of all the stars in our galaxy. This dataset contains data release 2 (DR2) products and consists of a catalogue of positions of stars, their parallax angle, proper motion, radial velocity and more.

XMM-Newton - Is the most sensitive X-ray observatory ever put in space. This dataset contains the spectral lightcurves of X-ray sources from the XMM archives. The light curves contain characteristic emission and absorption lines due to different chemical elements. Of particular interest is the identification of the Fe K_alpha line at 6.4 keV. 

Herschel - Is a retired far-infrared and sub-mm observatory. Herschel's primary objective was to study the formation of galaxies and the chemical composition of the solar system. In this dataset, we prepare images that could be used for identification of extended and galactic emission regions. 

These datasets can be used on Day 4's hackathon, but alternatively participants may wish to explore the vast data available on the ESA science archives, or bring along their own data from an ESA science mission.


Monday Dec 2

  • 09:00 - Bus from Hotel Exe moncloa to ESAC
  • 09:30 - Registration
  • 09:50 - Introduction
  • 10:00 - G. Belanger on Introduction to Statistics
  • 11:15 - Break
  • 11:45 - Tutorials (part 1)
  • 13:30 - Lunch
  • 14:45 - G. Belanger on Introduction to Machine Learning
  • 16:00 - Break
  • 16:30 - Tutorials (part 2)
  • 18:00 - Bus to Madrid

Tuesday Dec 3

  • 09:00 - Bus from Hotel Exe moncloa to ESAC
  • 10:00 - S. Raval on Clustering & Classification (part 1)
  • 11:15 - Break
  • 11:45 - S. Raval on Clustering & Classification (part 2)
  • 13:30 - Lunch
  • 14:45 - S. Raval on Clustering & Classification (part 3)
  • 16:00 - Break
  • 16:30 - Tutorials on Clustering & Classification 
  • 18:00 - Cocktail dinner in Hall of building D
  • 19:30 - Bus to Madrid

Wednesday Dec 4

  • 09:00 - Bus from Hotel Exe moncloa to ESAC
  • 10:00 - S. Raval on Parameter Estimation & Advanced Machine Learning Methods (part 1)
  • 11:15 - Break
  • 11:45 - S. Raval on Parameter Estimation & Advanced Machine Learning Methods (part 2)
  • 13:30 - Lunch
  • 14:45 - S. Raval on Parameter Estimation & Advanced Machine Learning Methods (part 3)
  • 16:00 - Break
  • 16:30 - Tutorials on Parameter Estimation & Advanced Machine Learning Methods
  • 18:00 - Bus to Madrid

Thursday Dec 5

  • 09:00 - Bus from Hotel Exe moncloa to ESAC
  • 10:00 - ESA science data archives hackathon (part 1)
  • 11:15 - Break
  • 11:45 - ESA science data archives hackathon (part 2) 
  • 13:30 - Lunch
  • 14:45 - Summary and Discussion
  • 16:00 - Break and END
  • 16:30 - Bus to Madrid

lectures & tutorials

The lecture and tutorial materials will be made available at the time of the workshop.


Python. Jupyter notebook. 


All lectures will be made available on the ESAC Data Analysis and Statistics YouTube channel


You can reach us at