16. Machine Learning with ESASky.


ESA supervisor: Guido de Marchi
Collaborator(s): Sandor Kruk, Marcos Lopez Caniego

Site: ESAC

The ESASky (sky.esa.int) portal hosts and serves a large number of data sets in HiPS  (Hierarchical Progressive Surveys) format from missions such as XMM-Newton, HST, JWST and Spitzer, among many others. These HiPS maps have been produced at different Healpix resolutions tesselating the sky in tens of thousands of small png or jpeg images that are used by ESASky to display the sky at different zoom levels. These HiPS files are a valuable resource for training machine learning models and are publicly available at the ESASky servers. However, accessing and navigating  these repositories is not straightforward and users would benefit from a dedicated mechanism to serve these images. The ESA Datalabs Science Exploitation platform already provides access to a wide variety of data sets, but not the HiPS files from ESASky. During this traineeship you will develop with us the mechanism to give users the possibility to access and use these images in ESA Datalabs. This will provide new valuable datasets for running ML experiments in ESA Datalabs and will ultimately enable new science.

Project duration: 6 months.

Desirable expertise or programming language:

  • Computer and data science. 
  • Programming experience in Python would be an advantage.


To apply for this project please fill in an online application form through the following link.