# Multimodal Universe: DESI Legacy Imaging Surveys North Galaxy Images (SSL Sample)

## Facts

- Provider: Multimodal Universe Collaboration (Polymathic AI, UniverseTBD)
- License: CC-BY-4.0 (https://www.legacysurvey.org/acknowledgment/)
- Size: 10,995 files
- Format: PNG, Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/UniverseTBD/mmu_ssl_legacysurvey_north
- Access: public, readable without an account
- Category: Astronomy & Space
- Source: https://arxiv.org/abs/2412.02527
- Page: https://scigantic.com/archive/a3316e5b-b803-5484-bd0c-914180d849ec
- Keywords: Astronomy & Space, spectroscopy, PNG, Parquet, Hugging Face

## About

Multi-band galaxy image cutouts from the northern footprint of the DESI Legacy Imaging Surveys, the self-supervised learning (SSL) galaxy sample, packaged as part of the Multimodal Universe, a collection of about 100 TB of machine-learning-ready astronomical data. The modality is imaging: each of 14,174,203 objects carries a g, r and z band flux image of 152 by 152 pixels at 0.262 arcsec per pixel with per-band PSF FWHM, together with catalogue photometry (flux_g, flux_r, flux_z, fiber fluxes, PSF depths), Galactic extinction E(B-V) and a spectroscopic redshift z_spec where one exists. The README lists 15 columns in 5,488 HEALPix partitions and 3.4 TiB on disk. Listing the repository (measured 2026-11-13) shows the main catalog folder mmu_ssl_legacysurvey_north/ (dataset/, hats.properties, partition_info.csv, point_map.fits), the margin catalog mmu_ssl_legacysurvey_north_10arcs/ and collection.properties; about 3.8 TB in all. Data are a HATS catalog: Apache Parquet partitioned by HEALPix tile, with a 10 arcsecond margin catalog for cross-matching. Open with LSDB (lsdb.open_catalog on the Hugging Face URL, with cone searches and crossmatch), or read the Parquet with pandas, pyarrow, dask or DuckDB. Uses include galaxy representation learning, photometric redshift estimation and morphology. Part of the Multimodal Universe (arXiv:2412.02527); released under CC-BY-4.0, cite the Multimodal Universe paper and the original survey.

## Load the data

```python
from datasets import load_dataset
ds = load_dataset("UniverseTBD/mmu_ssl_legacysurvey_north")
print(ds)
```

## Data structure

File types in sampled files: PNG, Parquet

Example file: `mmu_ssl_legacysurvey_north/dataset/Norder=4/Dir=0/Npix=1005.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/a3316e5b-b803-5484-bd0c-914180d849ec
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/a3316e5b-b803-5484-bd0c-914180d849ec
- Over MCP: `get_schema_card` with id `a3316e5b-b803-5484-bd0c-914180d849ec` at https://api.scigantic.com/mcp

## Related topics

- [Parquet datasets](https://scigantic.com/topics/parquet)
- [Astronomy & Space datasets](https://scigantic.com/topics/astronomy-and-space)
- [Spectroscopy datasets](https://scigantic.com/topics/spectroscopy)
- [PNG datasets](https://scigantic.com/topics/png)
