juliensimon/fireball-bolide-events

Fireball and Bolide Events

Credit: NASA/ESA

Part of a dataset collection on Hugging Face.

Dataset description

Atmospheric impact events (fireballs and bolides) detected by US government sensors, from NASA JPL CNEOS.

Fireballs are exceptionally bright meteors caused by small asteroids or large meteoroids entering the atmosphere at high speed. The largest events can release energy equivalent to tens or hundreds of kilotons of TNT.

The detection threshold for these sensors corresponds roughly to bolides with impact energies above 0.1 kilotons of TNT, equivalent to a meteoroid roughly 1 meter in diameter striking the atmosphere at typical velocities of 15-25 km/s. The largest event in the modern record is the Chelyabinsk airburst of February 2013, which released approximately 440 kilotons and produced a shockwave that injured over 1,500 people. Events of this magnitude (10-20 meter impactors) are estimated to occur roughly once per century, while meter-scale impacts happen several times per year.

The velocity components (Vx, Vy, Vz) in the Earth-Centered Earth-Fixed (ECEF) frame allow reconstruction of the pre-atmospheric orbit, connecting individual bolides to their parent populations in the asteroid or cometary reservoirs.

The geographic distribution of detected fireballs reflects both the true impact flux and sensor coverage biases. Most events are detected over open ocean or unpopulated regions, and coordinate data may be withheld for events near politically sensitive areas. The radiated energy represents only a fraction of the total kinetic energy, with the remainder partitioned into the shockwave, fragmentation, heating, and deceleration.

This dataset is suitable for tabular classification, tabular regression tasks.

Schema

Column	Type	Description	Sample	Null %
datetime	datetime64[us]	Date and time of peak brightness (UTC); derived from satellite and infrasound sensor data	1988-04-15 03:03:10	0.0%
radiated_energy_j	float64	Total optical energy radiated in joules (x10^10); proportional to kinetic energy; Chelyabinsk 2013: ~5x10^14 J; null for events without optical energy estimate	758.0	0.0%
impact_energy_kt	float64	Estimated total impact energy in kilotons of TNT equivalent; Chelyabinsk ~500 kt; null for ~half of events where conversion is not available	14.0	0.0%
latitude	float64	Geographic latitude of peak brightness in decimal degrees (positive = N, negative = S); null for events where coordinates are withheld or unavailable	-4.1	17.6%
lat_direction	str	Original source latitude hemisphere indicator: 'N' or 'S'; retained for provenance; use signed latitude for analysis	S	17.6%
longitude	float64	Geographic longitude of peak brightness in decimal degrees (positive = E, negative = W); null for events where coordinates are withheld or unavailable	124.3	17.6%
lon_direction	str	Original source longitude hemisphere indicator: 'E' or 'W'; retained for provenance; use signed longitude for analysis	E	17.6%
altitude_km	float64	Altitude at peak brightness in km; typical fireball burn-up range 20-80 km; null for events without altitude measurement	37.0	44.7%
velocity_kms	float64	Pre-entry velocity at peak brightness in km/s; range ~11 km/s (escape velocity minimum) to ~72 km/s (maximum relative to Earth's orbit); null for events without velocity measurement	17.1	66.5%

Quick stats

• 1,057 fireball events (1988-04-15 to 2026-04-08)
• 1057 events with measured impact energy
• 871 events with geographic coordinates
• Largest impact energy: 441.0 kt

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/fireball-bolide-events", split="train")
df = ds.to_pandas()

from datasets import load_dataset

ds = load_dataset("juliensimon/fireball-bolide-events", split="train")
df = ds.to_pandas()

# High-energy events (> 1 kiloton)
big = df[df["impact_energy_kt"] > 1].sort_values("impact_energy_kt", ascending=False)
print(big[["datetime", "impact_energy_kt", "latitude", "longitude"]])

# Plot events on a map
import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(12, 6))
coords = df.dropna(subset=["latitude", "longitude"])
ax.scatter(coords["longitude"], coords["latitude"],
           s=coords["impact_energy_kt"].fillna(0.1) * 5, alpha=0.5)
ax.set_xlabel("Longitude")
ax.set_ylabel("Latitude")
ax.set_title("Fireball Events by Location and Energy")
plt.show()

Data source

https://cneos.jpl.nasa.gov/fireballs/

Related datasets

• juliensimon/neo-close-approaches

• juliensimon/sentry-impact-risk

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{fireball_bolide_events,
  title = {Fireball and Bolide Events},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/fireball-bolide-events},
  publisher = {Hugging Face}
}

License

CC-BY-4.0