juliensimon/desi-dr1-redshifts

---
license: cc-by-4.0
pretty_name: "DESI DR1 Bright Galaxy Survey Redshifts"
language:
- en
description: "Spectroscopic redshifts from the Dark Energy Spectroscopic Instrument (DESI) Data Release 1 — Bright Galaxy Survey subset with reliable measurements (zwarn=0). The largest spectroscopic survey ever conducted."
task_categories:
- tabular-classification
- tabular-regression
tags:
- space
- galaxies
- redshifts
- desi
- spectroscopy
- cosmology
- astronomy
- open-data
- tabular-data
size_categories:
- 1M<n<10M
configs:
- config_name: default
data_files:
- split: train
path: data/desi-dr1-redshifts.parquet
default: true
---

# DESI DR1 Bright Galaxy Survey Redshifts

*Part of the [Astronomy Datasets](https://huggingface.co/collections/juliensimon/astronomy-datasets-69c24caf2f17e36128946743) collection on Hugging Face.*

Spectroscopic redshifts from the [Dark Energy Spectroscopic Instrument](https://www.desi.lbl.gov/)
(DESI) Data Release 1 — the **largest spectroscopic survey ever conducted**.
This dataset contains the **Bright Galaxy Survey (BGS)** subset: **5,000,000** sources
with reliable redshift measurements (`zwarn=0`, `main_primary=true`).

## Dataset description

DESI is a robotic fiber-fed spectrograph on the Mayall 4-meter telescope at
Kitt Peak National Observatory, capable of measuring 5,000 spectra simultaneously.
DR1 contains 28.4 million unique spectroscopic redshifts from 14,600+ square degrees.

The Bright Galaxy Survey targets galaxies with r < 19.5 magnitude during bright
lunar conditions. This curated subset includes only the main survey observations
with reliable redshift fits (zero warning flags) and primary measurements
(no duplicates).

**Breakdown by spectral type:**
- **3,519,828** galaxies (70.4%)
- **1,439,945** stars (28.8%)
- **40,227** QSOs (0.8%)

## Schema

| Column | Type | Description |
|--------|------|-------------|
| `targetid` | int64 | Unique DESI target identifier |
| `ra` | float64 | Right ascension — fiber position (degrees) |
| `dec` | float64 | Declination — fiber position (degrees) |
| `redshift` | float64 | Best-fit spectroscopic redshift |
| `redshift_err` | float64 | Redshift uncertainty |
| `spectype` | string | Spectral classification: GALAXY, STAR, or QSO |
| `subtype` | string | Spectral subtype (e.g., stellar type K, G, F) |
| `deltachi2` | float64 | Chi-squared difference between best and second-best fit |
| `chi2` | float64 | Best-fit chi-squared |
| `coadd_numexp` | int16 | Number of coadded exposures |
| `coadd_numnight` | int16 | Number of observation nights |
| `coadd_numtile` | int16 | Number of observed tiles |
| `coadd_exptime` | float32 | Total coadded exposure time (seconds) |

## Quick stats

- **5,000,000** sources with reliable redshifts
- Median redshift: **0.1658**
- Mean redshift: **0.1712**
- **3,519,828** galaxies, **1,439,945** stars, **40,227** QSOs
- Sky coverage: ~14,600 square degrees

## Usage

```python
from datasets import load_dataset

ds = load_dataset("juliensimon/desi-dr1-redshifts", split="train")
df = ds.to_pandas()

# Galaxy redshift distribution
galaxies = df[df["spectype"] == "GALAXY"]
print(f"{len(galaxies):,} galaxies, median z = {galaxies['redshift'].median():.4f}")

# Redshift histogram
import matplotlib.pyplot as plt
galaxies["redshift"].hist(bins=200, range=(0, 0.6))
plt.xlabel("Redshift")
plt.ylabel("Count")
plt.title("DESI BGS Galaxy Redshift Distribution")

# Sky coverage plot
plt.figure(figsize=(12, 6))
plt.scatter(df["ra"], df["dec"], s=0.01, alpha=0.1)
plt.xlabel("RA (deg)")
plt.ylabel("Dec (deg)")
plt.title("DESI DR1 BGS Sky Coverage")
```

## Data source

All data comes from the [DESI Data Release 1](https://data.desi.lbl.gov/doc/releases/dr1/)
via the [NOIRLab Astro Data Lab](https://datalab.noirlab.edu/) TAP service.

DESI Collaboration (2025). "The DESI Data Release 1." arXiv:2503.14745.

## Related datasets

- [sdss-dr18-spectra](https://huggingface.co/datasets/juliensimon/sdss-dr18-spectra) — SDSS DR18 optical spectroscopy
- [exoplanet-archive](https://huggingface.co/datasets/juliensimon/exoplanet-archive) — NASA Exoplanet Archive

## Pipeline

Source code: [juliensimon/space-datasets](https://github.com/juliensimon/space-datasets)

## Citation

```bibtex
@dataset{desi_dr1_redshifts,
author = {Simon, Julien},
title = {DESI DR1 Bright Galaxy Survey Redshifts},
year = {2026},
publisher = {Hugging Face},
url = {https://huggingface.co/datasets/juliensimon/desi-dr1-redshifts},
note = {Based on DESI DR1 (DESI Collaboration 2025) via NOIRLab Astro Data Lab}
}
```

## License

[CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)