Natural language processing of astronomical transients reports
2021-10-25, 10:15–10:30, Grand Ballroom

Nowadays, a wealth of astronomical data is easily accessible online from machine-readable resources, such as, e.g., source catalogs or experimental data repositories. Many tools exist to extract, process and combine information from them automatically. On the other hand, there is also a considerable amount of important information hidden in the natural language sources. Particularly interesting are the GCN Circulars and Astronomers Telegrams issued shortly after an interesting transient event occurs, such as a supernova, gravitational wave, or neutrino alert. They are widely used by the multi-messenger community to inform about those events and the results of their follow-up and to plan further observations. A considerable amount of time is spent by the researchers searching, reading, and comparing information from those reports. Here we present a tool-set based on modern natural language processing techniques that allow for automatic extraction and processing of information from the GCN Circulars and Astronomers Telegrams. Data such as source name and coordinates, flux level, detection or non-detection, etc., can now be automatically obtained. The python-base code is easily adaptable to the individual researcher or experiment needs.


Solutions for workflow management and reproducibility, Modernizing and maintaining telescopes, Building accessible and friendly user interfaces, Big data: How to deal with the 5 Vs (volume, velocity, variety, veracity, value)