---
license: other
license_name: cc-by-nc-3.0-igo
license_link: https://creativecommons.org/licenses/by-nc/3.0/igo/
pretty_name: "Gaia DR3 Binary Masses"
language:
  - en
description: "The Gaia DR3 binary masses catalog provides physical masses for binary star systems derived by the ESA Gaia mission. Masses are determined by combining astrometric non-single-star (NSS) solutions with"
task_categories:
  - tabular-regression
tags:
  - space
  - gaia
  - binary-stars
  - stellar-masses
  - esa
  - astronomy
  - open-data
  - tabular-data
  - parquet
size_categories:
  - 100K<n<1M
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/gaia_dr3_binary_masses.parquet
    default: true
---

# Gaia DR3 Binary Masses


<div align="center">
  <img src="banner.jpg" alt="Stellar field showing binary star systems" 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 binary masses catalog provides physical masses for binary star systems derived by the ESA Gaia mission. Masses are determined by combining astrometric non-single-star (NSS) solutions with spectroscopic radial velocities and photometric constraints, applying Kepler's third law to the orbital solutions.

These are among the most precisely measured stellar masses available from an all-sky survey, covering ~195,000 systems with primary masses ranging from ~0.1 to more than 10 solar masses. The combination_method column indicates which observational constraints (astrometry, spectroscopy, photometry) were used for each system.

The dataset is scientifically crucial for calibrating stellar evolution models, mass-luminosity relations, and testing stellar structure theory. Direct mass measurements from binary systems are the foundation of our understanding of how stars form, evolve, and die. Unlike masses estimated from stellar models or spectral fitting, these Gaia masses are derived geometrically from orbital dynamics, providing model-independent ground truth across a wide range of stellar types.

The flux ratio (F2/F1 in G-band) constrains the luminosity of the secondary star, enabling estimates of secondary radius and temperature when combined with the primary's known properties. Systems with measured m2 allow direct computation of the mass ratio q = m2/m1, a fundamental parameter in binary evolution theory that controls mass transfer rates, tidal circularization timescales, and ultimate system fate.


This dataset is suitable for **tabular regression** tasks.

## Schema

| Column | Type | Description | Sample | Null % |
|--------|------|-------------|--------|--------|
| `source_id` | int64 | Gaia DR3 source identifier of the primary star; use for cross-matching with other Gaia tables | 33711199137024 | 0.0% |
| `m1` | float64 | Primary star mass in solar masses (M☉), derived from orbital solution | 0.94867384 | 0.0% |
| `m1_lower` | float64 | Lower 1-sigma uncertainty on m1 (M☉) | 0.654218 | 0.0% |
| `m1_upper` | float64 | Upper 1-sigma uncertainty on m1 (M☉) | 0.9998373 | 0.0% |
| `m2` | float64 | Secondary component mass in solar masses (M☉); may be NaN if only primary mass is determined | 0.5688726 | 85.3% |
| `m2_lower` | float64 | Lower 1-sigma uncertainty on m2 (M☉) | 0.44135773 | 0.0% |
| `m2_upper` | float64 | Upper 1-sigma uncertainty on m2 (M☉) | 1.8444434 | 30.9% |
| `fluxratio` | float64 | Flux ratio of secondary to primary (F2/F1) in Gaia G-band | -0.026845658 | 90.1% |
| `fluxratio_lower` | float64 | Lower uncertainty on flux ratio | 0.0 | 0.2% |
| `fluxratio_upper` | float64 | Upper uncertainty on flux ratio | 0.59 | 33.0% |
| `combination_method` | str | Method used to derive masses: combination of astrometric, spectroscopic, and/or photometric constraints | Orbital+M1 | 0.0% |
| `m1_ref` | str | Reference for the primary mass estimate (short code) | IsocLum | 0.0% |
| `mass_ratio` | float64 | Mass ratio q = m2/m1; computed where both masses are available and m1 > 0 | 0.45398824218362815 | 85.3% |
| `m1_uncertainty` | float64 | Average 1-sigma uncertainty on m1 = (m1_upper + m1_lower) / 2.0 (M☉) | 0.82702765 | 0.0% |

## Quick stats

- **195,315** binary star systems
- Median primary mass (m1): 1.037 M☉
- Median secondary mass (m2): 0.578 M☉ (where available)
- Systems with measured m2: 28,706 (14.7%)
- Median mass ratio (m2/m1): 0.559

## Usage

```python
from datasets import load_dataset

ds = load_dataset("juliensimon/gaia-dr3-binary-masses", split="train")
df = ds.to_pandas()
```

```python
from datasets import load_dataset
import matplotlib.pyplot as plt

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

# Primary mass distribution
df["m1"].clip(upper=5).hist(bins=100, log=True)
plt.xlabel("Primary mass m1 (M☉)")
plt.ylabel("Count")
plt.title("Gaia DR3 Binary Star Primary Mass Distribution")
plt.show()

# m1 vs m2 scatter (systems with both measured)
has_m2 = df.dropna(subset=["m1", "m2"])
plt.hexbin(has_m2["m1"], has_m2["m2"], gridsize=80, mincnt=1, cmap="viridis")
plt.colorbar(label="Count")
plt.xlabel("Primary mass m1 (M☉)")
plt.ylabel("Secondary mass m2 (M☉)")
plt.title("Gaia DR3 Binary Mass Pairs")
plt.show()

# Mass ratio distribution
df["mass_ratio"].clip(0, 2).hist(bins=100)
plt.xlabel("Mass ratio q = m2/m1")
plt.ylabel("Count")
plt.title("Binary Mass Ratio Distribution")
plt.show()
```

## Data source

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

## Related datasets

- [juliensimon/gaia-dr3-spectroscopic-binaries](https://huggingface.co/datasets/juliensimon/gaia-dr3-spectroscopic-binaries)

- [juliensimon/gaia-dr3-eclipsing-binaries](https://huggingface.co/datasets/juliensimon/gaia-dr3-eclipsing-binaries)

- [juliensimon/wds](https://huggingface.co/datasets/juliensimon/wds)

> If you find this dataset useful, please consider [giving it a like](https://huggingface.co/datasets/juliensimon/gaia-dr3-binary-masses) on Hugging Face. It helps others discover it.

## About the author

Created by [Julien Simon](https://julien.org) — AI Operating Partner at Fortino Capital. Part of the [Space Datasets](https://julien.org/datasets) collection.

## Citation

```bibtex
@dataset{gaia_dr3_binary_masses,
  title = {Gaia DR3 Binary Masses},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/gaia-dr3-binary-masses},
  publisher = {Hugging Face}
}
```

## License


[CC-BY-NC-3.0-IGO](https://creativecommons.org/licenses/by-nc/3.0/igo/)