# Multimodal Universe: BTSbot ZTF Bright Transient Survey Alert Images

## Facts

- Provider: Multimodal Universe Collaboration (Polymathic AI, UniverseTBD)
- License: CC-BY-4.0 (https://zenodo.org/records/10839691)
- Size: 2,281 files
- Format: PNG, Parquet (in sampled files)
- Stored at: Hugging Face, https://huggingface.co/datasets/UniverseTBD/mmu_btsbot
- Access: public, readable without an account
- Category: Astronomy & Space
- Source: https://arxiv.org/abs/2401.15167
- Page: https://scigantic.com/archive/75dd9eb1-bd36-5254-b022-e4262fa67973
- Keywords: Astronomy & Space, PNG, Parquet, Hugging Face

## About

Zwicky Transient Facility (ZTF) alert images and alert-packet features from the training set of BTSbot, a multimodal convolutional neural network that automates source selection for the ZTF Bright Transient Survey (Rehemtulla et al. 2024, arXiv:2401.15167), packaged as part of the Multimodal Universe collection of machine-learning-ready astronomical data. The modality is time-domain imaging with tabular metadata: each of 818,214 alerts carries image cutouts by band and view (science, reference and difference images), together with photometry (magpsf, sigmapsf, magap), real-bogus scores (drb), star-galaxy scores and distances to nearby Pan-STARRS and Gaia sources, light-curve summaries such as peakmag and days_since_peak, and labels such as is_SN. The README lists 92 columns in 2,078 HEALPix partitions and 18.9 GiB on disk. Listing the repository (measured 2026-11-13) shows the main catalog folder mmu_btsbot/ (dataset/, hats.properties, partition_info.csv, point_map.fits), the margin catalog mmu_btsbot_10arcs/ and collection.properties; about 21 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. 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_btsbot")
print(ds)
```

## Data structure

File types in sampled files: PNG, Parquet

Example file: `mmu_btsbot/dataset/Norder=10/Dir=10000/Npix=18936.parquet`

## Open it

- In a hosted notebook with the data mounted: https://scigantic.com/archive/75dd9eb1-bd36-5254-b022-e4262fa67973
- Full record as JSON (schema card included): https://api.scigantic.com/api/archives/75dd9eb1-bd36-5254-b022-e4262fa67973
- Over MCP: `get_schema_card` with id `75dd9eb1-bd36-5254-b022-e4262fa67973` 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)
- [PNG datasets](https://scigantic.com/topics/png)
