# Multimodal Universe: Chandra Source Catalog 2.1 X-ray Spectra (128,900 Sources, HATS)

## Facts

- Provider: Multimodal Universe Collaboration (Polymathic AI, UniverseTBD); Chandra X-ray Center (SAO/NASA)
- License: Public Domain (https://asc.cfa.harvard.edu/cda/public.html)
- Size: 2,717 files
- Format: PNG, Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/UniverseTBD/mmu_chandra_spectra
- Access: public, readable without an account
- Category: Astronomy & Space
- Source: https://arxiv.org/abs/2412.02527
- Page: https://scigantic.com/archive/59b37261-90b7-507f-b140-9f292e84eff7
- Keywords: Astronomy & Space, spectroscopy, PNG, Parquet, Hugging Face

## About

X-ray spectra of point and extended sources from the Chandra Source Catalog version 2.1 (CSC 2.1), built from Chandra X-ray Observatory ACIS event data, packaged as part of the Multimodal Universe collection of machine-learning-ready astronomical data. The modality is 1D X-ray spectroscopy: each of 128,900 sources has a binned spectrum over 0.5 to 8 keV (energy bin centre, low and high edges, count-rate flux and flux error) plus catalogue features from CSC: broad-band aperture flux, black-body aperture flux, flux significance, the three hardness ratios hard_hm, hard_hs and hard_ms, and the variability index and probability. Sources with fewer than 40 counts or S/N below 4 are excluded. Objects include X-ray binaries, active galactic nuclei, active stars and supernova remnant knots. The README lists 13 columns in 2,257 HEALPix partitions and 276.5 MiB on disk; about 196 MB stored (measured 2026-11-16), main catalog mmu_chandra_spectra/ and margin mmu_chandra_spectra_10arcs/. 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 X-ray source classification and cross-modal learning with optical spectra. The Chandra X-ray Center places released Chandra observations in the public domain; cite the CSC (Evans et al. 2010, doi:10.25574/csc2) and the Multimodal Universe paper.

## Load the data

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

## Data structure

File types in sampled files: PNG, Parquet

Example file: `mmu_chandra_spectra/dataset/Norder=10/Dir=10010000/Npix=10019805.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/59b37261-90b7-507f-b140-9f292e84eff7
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/59b37261-90b7-507f-b140-9f292e84eff7
- Over MCP: `get_schema_card` with id `59b37261-90b7-507f-b140-9f292e84eff7` 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)
