The dataset comes from the Kaggle competition "Photometric LSST Astronomical Time Series Classification Challenge" (PLAsTiCC) [1]. Held in late 2018, the competition was in preparation for first light of the Large Synoptic Survey Telescope (LSST) [2].
LSST endeavors to detect transients - stars that are actively changing - and notify astronomers of their presence. Examples of transients are:
Each night LSST will capture 20 Terabytes of data and and classify about 2.5 billion stars and galaxies. The goal is:
It is a huge classification challenge!
References: