CoolFace
Datasetpublic

juliensimon/gaia-dr3-compact-companions

Gaia DR3 Compact Companion Candidates Credit: NASA/JPL-Caltech Part of a dataset collection on Hugging Face. Dataset description The Gaia DR3 Compact Companion Candidates catalog contains ~6,300 candidates for binary systems where a normal (luminous) star orbits an unseen compact object — a white dwarf, neutron star, or black hole. These candidates were identified by the Gaia DR3 variability processing pipeline through the detection of ellipsoidal… See the full description on the dataset page: https://huggingface.co/datasets/juliensimon/gaia-dr3-compact-companions.

sourceHugging Faceotherupdated 7d agoView on Hugging Face
0likes50downloads
Dataset Card

Gaia DR3 Compact Companion Candidates

<div align="center"> <img src="banner.jpg" alt="Artist's concept of a black hole binary system" width="400"> <p><em>Credit: NASA/JPL-Caltech</em></p> </div>

Part of a [dataset collection](https://huggingface.co/collections/juliensimon/astronomy-datasets-69c24caf2f17e36128946743) on Hugging Face.

Dataset description

The Gaia DR3 Compact Companion Candidates catalog contains ~6,300 candidates for binary systems where a normal (luminous) star orbits an unseen compact object — a white dwarf, neutron star, or black hole. These candidates were identified by the Gaia DR3 variability processing pipeline through the detection of ellipsoidal variations in the primary star's light curve.

Ellipsoidal variability arises when the primary star is tidally distorted into a prolate ellipsoid by the gravitational pull of its compact companion. As the system orbits, the projected cross-section of the distorted star changes, producing a characteristic double-humped light curve at half the true orbital period. Because the compact companion emits negligible light compared to the primary, no eclipse is required — the signal is purely photometric, making this technique uniquely sensitive to quiescent (non-accreting) compact objects that are invisible to X-ray telescopes.

The minimum mass ratio columns (modminmassratio, modminmassratioonesigma, modminmassratiothree_sigma) constrain the companion mass assuming a range of orbital inclinations. A 3-sigma minimum mass ratio exceeding ~0.6 means the companion is too massive to be a main-sequence star at any plausible inclination, strongly suggesting a degenerate remnant. The alpha parameter encodes the ellipsoidal amplitude, which depends on the tidal distortion strength and therefore on the companion-to-primary mass ratio and the orbital separation.

This catalog represents one of the largest systematic searches for quiescent black hole and neutron star binaries ever conducted, and directly complements X-ray binary catalogs (HMXBs, LMXBs) which only detect systems currently undergoing active accretion. Identifying the dormant population is essential for constraining the true space density of stellar-mass black holes and the rate of compact object formation from stellar evolution.

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

Schema

ColumnTypeDescriptionSampleNull %
source_idInt64Gaia DR3 unique source identifier; use for cross-matching with other Gaia tables362263655149305600.0%
periodfloat64Orbital period in days; derived from ellipsoidal variability modeling0.25583030974971060.0%
period_errorfloat64Uncertainty on the orbital period (days)5.702947e-050.0%
t0_gfloat64Reference epoch (time of maximum in G-band light curve) expressed as BJD - 2455197.5 days1975.97952461548520.0%
t0_g_errorfloat64Uncertainty on t0_g (days)0.00059514320.0%
t0_bpfloat64Reference epoch (time of maximum in BP-band light curve) expressed as BJD - 2455197.5 days1975.97708133128964.6%
t0_bp_errorfloat64Uncertainty on t0_bp (days)0.00410315774.6%
t0_rpfloat64Reference epoch (time of maximum in RP-band light curve) expressed as BJD - 2455197.5 days1975.97758272570564.7%
t0_rp_errorfloat64Uncertainty on t0_rp (days)0.00662079974.7%
model_mean_gfloat64Mean G-band magnitude derived from the fitted ellipsoidal variability model18.4608230.0%
model_mean_g_errorfloat64Uncertainty on the model mean G magnitude0.00176676930.0%
model_mean_bpfloat64Mean BP-band magnitude derived from the fitted ellipsoidal variability model19.0265274.6%
model_mean_bp_errorfloat64Uncertainty on the model mean BP magnitude0.0354401064.6%
model_mean_rpfloat64Mean RP-band magnitude derived from the fitted ellipsoidal variability model17.6472824.7%
model_mean_rp_errorfloat64Uncertainty on the model mean RP magnitude0.0205278044.7%
mod_min_mass_ratiofloat64Minimum companion-to-primary mass ratio (Mcompanion / Mprimary) under median inclination assumption; lower bound on companion mass1.35253180.0%
mod_min_mass_ratio_one_sigmafloat64Minimum mass ratio at 1-sigma upper confidence; accounts for inclination uncertainty0.99880510.0%
mod_min_mass_ratio_three_sigmafloat64Minimum mass ratio at 3-sigma upper confidence; values > ~0.6 indicate the companion cannot be an ordinary main-sequence star, suggesting a compact object (white dwarf, neutron star, or black hole)0.539819540.0%
alphafloat64Ellipsoidal variability amplitude parameter; measures the degree of tidal deformation of the primary star by the unseen compact companion; higher values indicate stronger tidal distortion1.30.0%
bp_rpfloat64BP minus RP color index (modelmeanbp - modelmeanrp); traces stellar temperature and reddening1.3792450000000015.0%
likely_compactboolBoolean flag: True if modminmassratiothree_sigma > 0.5, indicating a likely compact (sub-stellar or degenerate) companionTrue0.0%

Quick stats

  • —6,306 compact companion candidates
  • —319 likely compact objects (3-sigma mass ratio > 0.5)
  • —Median orbital period: 0.4152 days (range: 0.250 – 2.5 days)
  • —Median minimum mass ratio: 0.7019

Usage

python
from datasets import load_dataset

ds = load_dataset("juliensimon/gaia-dr3-compact-companions", split="train")
df = ds.to_pandas()
python
from datasets import load_dataset

ds = load_dataset("juliensimon/gaia-dr3-compact-companions", split="train")
df = ds.to_pandas()

# Likely compact companions (neutron star / black hole candidates)
compact = df[df["likely_compact"] == True]
print(f"Likely compact companions: {len(compact):,}")

# Period vs mass ratio scatter (highlight compact candidates)
import matplotlib.pyplot as plt
fig, ax = plt.subplots(figsize=(10, 6))
mask = df["likely_compact"]
ax.scatter(df.loc[~mask, "period"], df.loc[~mask, "mod_min_mass_ratio_three_sigma"],
           s=5, alpha=0.4, label="Other", color="steelblue")
ax.scatter(df.loc[mask, "period"], df.loc[mask, "mod_min_mass_ratio_three_sigma"],
           s=10, alpha=0.7, label="Likely compact", color="crimson")
ax.set_xlabel("Period (days)")
ax.set_ylabel("Min mass ratio (3σ)")
ax.set_xscale("log")
ax.legend()
plt.title("Gaia DR3 Compact Companion Candidates")
plt.show()

# Period histogram
df["period"].hist(bins=100, log=True)
plt.xlabel("Period (days)")
plt.ylabel("Count")
plt.title("Orbital Period Distribution")
plt.show()

Data source

https://gea.esac.esa.int/archive/

Related datasets

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

bibtex
@dataset{gaia_dr3_compact_companions,
  title = {Gaia DR3 Compact Companion Candidates},
  author = {Simon, Julien},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/gaia-dr3-compact-companions},
  publisher = {Hugging Face},
  note = {Derived from ESA Gaia Archive, https://gea.esac.esa.int/archive/}
}

License

CC-BY-NC-3.0-IGO