# Multimodal Universe: DESI PROVABGS Galaxy Stellar Mass and Star Formation Posteriors

## Facts

- Provider: Multimodal Universe Collaboration (Polymathic AI, UniverseTBD)
- License: CC-BY-4.0 (https://data.desi.lbl.gov/doc/acknowledgments/)
- Size: 164 files
- Format: PNG, Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/UniverseTBD/mmu_desi_provabgs
- Access: public, readable without an account
- Category: Astronomy & Space
- Source: https://arxiv.org/abs/2208.08512
- Page: https://scigantic.com/archive/06208409-46f8-5ad4-b5f7-36a6ebe9e197
- Keywords: Astronomy & Space, spectroscopy, PNG, Parquet, Hugging Face

## About

Galaxy physical properties from PROVABGS (PRObabilistic Value-Added Bright Galaxy Survey; Hahn et al., arXiv:2208.08512), Bayesian spectral energy distribution (SED) fits to DESI Bright Galaxy Survey (BGS) spectra and photometry, packaged as part of the Multimodal Universe collection of machine-learning-ready astronomical data. The modality is a catalogue of derived properties: each of 222,752 galaxies carries 100 posterior MCMC samples of 13 SED parameters and the best-fit parameter vector, log stellar mass, mass-weighted metallicity and age, average star formation rate, redshift, g, r, z and W1 magnitudes and the survey completeness weights. The README lists 22 columns in 73 HEALPix partitions and 1.3 GiB on disk. Listing the repository (measured 2026-11-13) shows the main catalog folder mmu_desi_provabgs/ (dataset/, hats.properties, partition_info.csv, point_map.fits), the margin catalog mmu_desi_provabgs_10arcs/ and collection.properties; about 1.3 GB 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. It serves as regression targets for models trained on DESI spectra or Legacy Surveys images. 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_desi_provabgs")
print(ds)
```

## Data structure

File types in sampled files: PNG, Parquet

Example file: `mmu_desi_provabgs/dataset/Norder=4/Dir=0/Npix=1002.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/06208409-46f8-5ad4-b5f7-36a6ebe9e197
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/06208409-46f8-5ad4-b5f7-36a6ebe9e197
- Over MCP: `get_schema_card` with id `06208409-46f8-5ad4-b5f7-36a6ebe9e197` 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)
