# Multimodal Universe: TESS SPOC Light Curves for 1.1 Million Stars (HATS)

## Facts

- Provider: Multimodal Universe Collaboration (Polymathic AI, UniverseTBD); TESS SPOC (NASA Ames) via MAST
- License: CC0-1.0 (https://archive.stsci.edu/publishing/data-use)
- Size: 2,158 files
- Format: PNG, Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/UniverseTBD/mmu_tess_spoc
- Access: public, readable without an account
- Category: Astronomy & Space
- Source: https://arxiv.org/abs/2412.02527
- Page: https://scigantic.com/archive/c96375c8-cdb5-5ac8-bfae-b5d9a3885818
- Keywords: Astronomy & Space, time series, PNG, Parquet, Hugging Face

## About

Photometric light curves from NASA's Transiting Exoplanet Survey Satellite (TESS), produced by the Science Processing Operations Center (SPOC) pipeline at NASA Ames and distributed by MAST, packaged as part of the Multimodal Universe collection of machine-learning-ready astronomical data. The modality is time series: each of 1,122,883 stars carries a nested lightcurve with time in BTJD (barycentric TESS Julian date), PDC_SAP flux and flux error in electrons per second, plus ra, dec and object_id (the TIC identifier). Stars are observed for about 27 days per sector at 20 second to 30 minute cadence; only selected sectors are included and points are cleaned with the SPOC quality flags. The README lists 4 data columns in 1,084 HEALPix partitions and 64.4 GiB on disk; the repository uses about 70 GB (measured 2026-11-16), with main catalog mmu_tess_spoc/ and margin catalog mmu_tess_spoc_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 exoplanet transit search, variable star and eclipsing binary classification, stellar parameter regression and light-curve foundation models. MAST states that NASA mission data it hosts are licensed under CC0; acknowledge TESS and MAST and cite the Multimodal Universe paper (arXiv:2412.02527).

## Load the data

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

## Data structure

File types in sampled files: PNG, Parquet

Example file: `mmu_tess_spoc/dataset/Norder=10/Dir=4990000/Npix=4993020.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/c96375c8-cdb5-5ac8-bfae-b5d9a3885818
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/c96375c8-cdb5-5ac8-bfae-b5d9a3885818
- Over MCP: `get_schema_card` with id `c96375c8-cdb5-5ac8-bfae-b5d9a3885818` at https://api.scigantic.com/mcp

## Related topics

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