juliensimon/gamma-ray-bursts

Fermi GBM Gamma-Ray Burst Catalog

Credit: NASA/DOE/Fermi LAT Collaboration

Part of a dataset collection on Hugging Face.

Dataset description

GRB detections from Fermi Gamma-ray Burst Monitor with duration, flux, and spectral parameters.

Gamma-ray bursts (GRBs) are the most energetic explosions in the universe. They come in two classes based on duration: short GRBs (T90 < 2 s, from neutron star mergers) and long GRBs (T90 >= 2 s, from massive star collapse). The Fermi GBM has been detecting GRBs across the full unocculted sky since 2008.

This dataset includes duration measurements (T90, T50), fluence, peak flux, and Band function spectral parameters for each burst.

The physical dichotomy between short and long GRBs reflects fundamentally different progenitor systems. Short GRBs (T90 < 2 s) arise from the coalescence of compact binary systems -- neutron star-neutron star or neutron star-black hole mergers -- as spectacularly confirmed by the joint gravitational-wave and electromagnetic detection of GRB 170817A / GW170817. Long GRBs (T90 > 2 s) are produced by the core collapse of massive Wolf-Rayet stars, where a newly formed black hole launches ultra-relativistic jets that punch through the stellar envelope.

The Fermi GBM is one of the workhorses of modern GRB astronomy, detecting roughly 240 bursts per year across its 12 sodium iodide (NaI, 8 keV - 1 MeV) and 2 bismuth germanate (BGO, 200 keV - 40 MeV) detectors. This catalog is a cornerstone for multi-messenger astrophysics: GBM triggers initiate rapid follow-up campaigns across the electromagnetic spectrum and provide temporal coincidence windows for searches in gravitational-wave and neutrino data.

This dataset is suitable for tabular classification tasks.

Schema

Column	Type	Description	Sample	Null %
name	str	GRB designation encoding the discovery date and sequence letter (e.g. 'GRB 170817A' = Aug 17 2017, first event that day); 'GRB 170817A' is the electromagnetic counterpart to GW170817, the first confirmed neutron-star merger	GRB260421351	0.0%
trigger_time	datetime64[ns]	UTC time of the GBM on-board trigger; converted from Modified Julian Date; precision ~milliseconds	2026-04-21 08:25:19.590643277	0.0%
ra	float64	Right ascension of best-fit GRB position (ICRS J2000.0, degrees, 0-360); initial GBM localization uncertainty is several degrees; null if localization failed	294.97	0.0%
dec	float64	Declination of best-fit GRB position (ICRS J2000.0, degrees, -90 to +90); null if localization failed	64.15	0.0%
t90	float64	Duration containing 90% of the burst's total photon counts (seconds); short GRBs have T90 < 2 s (neutron star mergers), long GRBs have T90 > 2 s (massive star collapse); bimodal distribution separates two physically distinct progenitor populations; null for bursts with insufficient counts	19.2003	0.0%
t90_error	float64	1-sigma statistical uncertainty on T90 (seconds); null when T90 is not measured	1.14489	0.0%
t50	float64	Duration containing the central 50% of burst counts (seconds); narrower than T90, less sensitive to faint extended emission; useful for comparing burst timescales across detectors	8.96015	0.0%
t50_error	float64	1-sigma statistical uncertainty on T50 (seconds); null when T50 is not measured	0.572456	0.0%
fluence	float64	Total gamma-ray fluence integrated over the burst duration (erg/cm^2); proxy for apparent isotropic energy release; null when spectral fit did not converge	6.09021e-06	0.0%
fluence_error	float64	1-sigma uncertainty on fluence (erg/cm^2); null when fluence is not measured	3.46374e-08	0.0%
flux_256	float64	Peak photon flux measured on a 256 ms timescale (photons/cm^2/s); determines detectability and is used in the logN-logP distribution; null when peak flux measurement failed	9.99223	0.0%
pflx_best_fitting_model	str	Name of the spectral model providing the best fit to the peak-flux time interval (e.g. 'band', 'comp', 'plaw', 'sbpl'); drives which set of spectral parameters is most reliable	pflx_comp	16.5%
flnc_band_ampl	float64	Amplitude (normalization) of the Band function fit to the time-integrated (fluence) spectrum (photons/cm^2/s/keV at pivot energy); null when the Band model is not the best fit or fit failed	0.015812058	15.2%
flnc_band_epeak	float64	Peak energy of the nuFnu spectrum from the Band function fit (keV); most GRBs fall between 100-2000 keV; correlates with isotropic luminosity (Amati relation); null when Band fit failed	111.80244	15.2%
flnc_band_alpha	float64	Low-energy photon spectral index of the Band function (dimensionless); typically -1.5 to 0; values harder than -2/3 violate synchrotron line-of-death, constraining emission models; null when Band fit failed	-1.2063167	15.2%
flnc_band_beta	float64	High-energy photon spectral index of the Band function (dimensionless); typically -3 to -2; describes the steep spectral cutoff above E_peak; null when Band fit failed or high-energy data insufficient	-2.0000999	15.2%
duration_class	str	Physical classification by T90: 'short' (T90 < 2 s, compact binary merger progenitor) or 'long' (T90 >= 2 s, massive star core collapse progenitor); null when T90 is unavailable	long	0.0%

Quick stats

• 4,269 gamma-ray bursts
• 713 short GRBs, 3,555 long GRBs
• Date range: 2008-07-14 to 2026-04-21
• Brightest burst: GRB221009553 (fluence 3.88e-02 erg/cm^2)

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/gamma-ray-bursts", split="train")
df = ds.to_pandas()

from datasets import load_dataset

ds = load_dataset("juliensimon/gamma-ray-bursts", split="train")
df = ds.to_pandas()

# Short vs long GRBs
short = df[df["duration_class"] == "short"]
long = df[df["duration_class"] == "long"]
print(f"{len(short):,} short, {len(long):,} long GRBs")

# Brightest bursts
top = df.nlargest(10, "fluence")[["name", "trigger_time", "fluence", "t90"]]

# T90 distribution
import matplotlib.pyplot as plt
df["t90"].dropna().apply(lambda x: max(x, 1e-3)).hist(bins=50, log=True)
plt.xlabel("T90 (s)")
plt.title("GRB Duration Distribution")
plt.show()

Data source

https://heasarc.gsfc.nasa.gov/W3Browse/fermi/fermigbrst.html

Related datasets

• juliensimon/pulsar-catalog

• juliensimon/supernova-remnants

• juliensimon/gravitational-wave-events

If you find this dataset useful, please consider giving it a like on Hugging Face. It helps others discover it.

About the author

Created by Julien Simon — AI Operating Partner at Fortino Capital. Part of the Space Datasets collection.

Citation

@dataset{gamma_ray_bursts,
  title = {Fermi GBM Gamma-Ray Burst Catalog},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/gamma-ray-bursts},
  publisher = {Hugging Face}
}

License

CC-BY-4.0