The Gaia Utility for the Analysis of Self-Organizing Maps (GUASOM) is an analysis tool that allows segmenting the information present in any other massive spectroscopic or spectrophotometric database. Self-Organizing Map is a non-supervised technique based on competitive neural networks, GUASOM displays these maps alongside additional information. The SOM is trained using a Python tool available in SPACIOUS and, once processed, the maps can be sent to GUASOM for visualisation in a web browser.

The figure above demonstrates GUASOM’s ability to group Gaia Data Release 3 (GDR3) data, in this case with 163919 BP/RP spectra. This data is accessible from the SPACIOUS platform, where a tutorial will be available to train and visualise this map.