license: apache-2.0
size_categories:
- 1B<n
task_categories:
- tabular-classification
- feature-extraction
tags:
- astronomy
- astrophysics
- gaia-dr3
- stellar-data
- space
Gaia-DR3: A Billion-Star Dataset for Galactic Analysis
This dataset is a high-fidelity, pre-processed collection of over 1 Billion stellar records derived from the European Space Agency (ESA) Gaia Mission (Data Release 3). It is specifically curated for large-scale galactic archaeology, 3D mapping, and training advanced machine learning models like SADIM-54M.
1. Dataset Overview
The dataset provides a comprehensive snapshot of the Milky Way, covering astrometric, kinematic, and photometric parameters. It has been optimized for high-performance computing and AI-driven astronomical frameworks.
2. Feature Schema (13 Core Parameters)
The dataset is structured with 13 essential features for understanding stellar dynamics:
| Feature Name | Data Type | Scientific Description |
|---|---|---|
| source_id | int64 |
Unique Gaia DR3 identifier for each stellar source. |
| ra / dec | float64 |
Equatorial Coordinates: Right Ascension & Declination. |
| l / b | float64 |
Galactic Coordinates: Longitude & Latitude relative to the Galaxy. |
| pmra / pmdec | float64 |
Proper Motion: Angular velocity of the star across the sky (mas/yr). |
| d_pc | float64 |
Distance: Calculated distance from Earth in Parsecs ($1/parallax$). |
| x, y, z | float64 |
3D Cartesian: Heliocentric position relative to the Sun. |
| abs_m | float64 |
Absolute Magnitude: The intrinsic brightness of the star. |
| bp_rp | float32 |
Color Index: Difference between BP and RP (Temperature indicator). |
3. Usage & Access (Streaming Mode)
Note: Due to the massive scale of this dataset (1B+ rows), downloading the full files to a local machine is not recommended. Use the Streaming Mode provided by the Hugging Face datasets library to process data on the fly:
from datasets import load_dataset
# Stream the dataset directly without downloading the full files
dataset = load_dataset("samfatnassi/gaia-dr3", split="train", streaming=True)
# Access a single stellar record
star_record = next(iter(dataset))
print(star_record)
4. Integration with SADIM-54M
This dataset serves as the foundational "knowledge base" for the SADIM-54M Model. While the dataset provides the raw observational facts, the model provides the analytical intelligence to predict and classify these stars.
5. Research & Ethics (Open Science)
This dataset is released under the Apache 2.0 License. It is provided as a contribution to Open Science and Humanity, encouraging researchers, students, and developers worldwide to explore the mysteries of our galaxy without boundaries.
Data Source: European Space Agency (ESA) Gaia Mission
Project Lead: KilmaAI / Sadim