High School Student Mines NASA Archive to Map 1.5 Million Variable Objects

A machine-learning pipeline reprocessed NASA’s NEOWISE infrared archive, turning nearly 200 billion measurements into a catalog of previously uncatalogued variable sources.

Overview

Pasadena High School student Matteo Paz identified about 1.5 million previously uncatalogued variable objects by analyzing WISE/NEOWISE infrared data.
He built an automated system in six weeks to scan the full archive of roughly 200 billion single-exposure measurements and track changes in brightness over time.
The model flags faint flickering, pulsing and fading patterns associated with quasars, binary stars, supernovae and young stars obscured by dust.
Paz collaborated with Caltech/IPAC researchers under Davy Kirkpatrick’s guidance, with the results reviewed by scientists and published in The Astronomical Journal.
The project earned Paz the $250,000 top Regeneron Science Talent Search prize and provides a rich resource for future studies and follow-up observations.