juliensimon/first-radio-catalog

---
license: cc-by-4.0
pretty_name: "FIRST Radio Survey Catalog"
language:
- en
description: "The Faint Images of the Radio Sky at Twenty-cm (FIRST) survey catalog, covering 10,575 square degrees at 1.4 GHz with 5 arcsecond resolution using the NRAO VLA. The FIRST survey used the VLA in its B"
task_categories:
- tabular-classification
- tabular-regression
tags:
- space
- radio
- first
- vla
- astronomy
- 1400mhz
- open-data
- tabular-data
- parquet
size_categories:
- 100K<n<1M
configs:
- config_name: default
data_files:
- split: train
path: data/first_radio_sources.parquet
default: true
---

# FIRST Radio Survey Catalog

<div align="center">
<img src="banner.jpg" alt="Deep Space Network antenna at Goldstone" width="400">
<p><em>Credit: NASA/JPL-Caltech</em></p>
</div>

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

## Dataset description

The Faint Images of the Radio Sky at Twenty-cm (FIRST) survey catalog, covering 10,575 square degrees
at 1.4 GHz with 5 arcsecond resolution using the NRAO VLA.

The FIRST survey used the VLA in its B-configuration to produce a map of the radio sky at
1.4 GHz with ~5" resolution and a typical rms of 0.15 mJy/beam. The catalog includes
source positions, peak and integrated flux densities, and fitted source sizes.

The FIRST survey was designed as a radio counterpart to the Palomar Observatory Sky Survey, targeting the north and south Galactic caps where optical and infrared surveys provide the richest multi-wavelength context. Its combination of sub-arcsecond positional accuracy and milliJansky sensitivity makes it ideally suited for identifying the radio counterparts of optically selected quasars, galaxies, and other extragalactic objects.

Unlike the broader but lower-resolution NVSS, FIRST resolves the internal structure of extended radio sources, revealing jets, lobes, and hotspots in radio galaxies. The survey's 5-arcsecond beam allows morphological classification of sources and reliable separation of core-dominated and lobe-dominated radio AGN.

## Schema

| Column | Type | Description | Sample | Null % |
|--------|------|-------------|--------|--------|
| `source_name` | object | FIRST source identifier in the format "JHHMMSS.s+DDMMSS" derived from J2000 position | J030004.5+000946 | 0.0% |
| `ra_deg` | float64 | ICRS J2000.0 right ascension in degrees (0-360); positional accuracy ~0.5 arcsec | 45.01904166666666 | 0.0% |
| `dec_deg` | float64 | ICRS J2000.0 declination in degrees; survey covers ~-10 to +62 deg | 0.1630138888888888 | 0.0% |
| `peak_flux_mjy` | float64 | Peak surface brightness at 1.4 GHz in mJy/beam; detection threshold ~0.75 mJy/beam | 1.04 | 0.0% |
| `integrated_flux_mjy` | float64 | Total integrated flux density from Gaussian fit in mJy; use for luminosity calculations | 0.97 | 0.0% |
| `rms_mjy` | float64 | Local rms noise at source position in mJy/beam; typically ~0.15 mJy/beam | 0.104 | 0.0% |
| `major_axis_arcsec` | float64 | Fitted (convolved) major-axis FWHM in arcsec; includes 5-arcsec beam | 0.39 | 0.0% |
| `minor_axis_arcsec` | float64 | Fitted (convolved) minor-axis FWHM in arcsec | 0.0 | 0.0% |
| `position_angle_deg` | float64 | Fitted position angle of major axis in degrees east from north (0-180) | 119.1 | 0.0% |
| `deconv_major_arcsec` | float64 | Deconvolved major axis in arcsec; 0 for unresolved point sources | 6.17 | 0.0% |
| `deconv_minor_arcsec` | float64 | Deconvolved minor axis in arcsec; 0 for unresolved sources | 5.28 | 0.0% |
| `deconv_pa_deg` | float64 | Deconvolved position angle in degrees; meaningful only when deconv_major_arcsec > 0 | 169.1 | 0.0% |
| `is_resolved` | bool | Derived: True if deconv_major_arcsec > 0 (source resolved above 5-arcsec beam) | True | 0.0% |

## Quick stats

- **946,432** radio sources at 1.4 GHz
- **946,432** resolved sources (100.0% of catalog)
- Median peak flux: **2.09** mJy/beam
- Sky coverage: 10,575 square degrees (north + south Galactic caps)
- Angular resolution: 5 arcsec (VLA B-configuration)

## Usage

```python
from datasets import load_dataset

ds = load_dataset("juliensimon/first-radio-catalog", split="train")
df = ds.to_pandas()

# Flux distribution (log scale — most sources are faint)
import matplotlib.pyplot as plt

fig, axes = plt.subplots(1, 2, figsize=(14, 5))

df["peak_flux_mjy"].clip(upper=100).hist(bins=200, log=True, ax=axes[0])
axes[0].set_xlabel("Peak flux (mJy/beam)")
axes[0].set_ylabel("Count")
axes[0].set_title("FIRST Source Flux Distribution")

# Sky coverage map
axes[1].scatter(df["ra_deg"], df["dec_deg"], s=0.01, alpha=0.1)
axes[1].set_xlabel("RA (deg)")
axes[1].set_ylabel("Dec (deg)")
axes[1].set_title("FIRST Survey Sky Coverage")
plt.tight_layout()
plt.show()

# Resolved vs unresolved statistics
print(f"Resolved: {df['is_resolved'].sum():,}")
print(f"Unresolved: {(~df['is_resolved']).sum():,}")
```

## Data source

https://vizier.cds.unistra.fr/viz-bin/VizieR-3?-source=VIII/92

## Citation

```bibtex
@dataset{first_radio_catalog,
title = {FIRST Radio Survey Catalog},
author = {juliensimon},
year = {2026},
url = {https://huggingface.co/datasets/juliensimon/first-radio-catalog},
publisher = {Hugging Face}
}
```

## License

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