juliensimon/henry-draper-catalog

Henry Draper Catalogue

Credit: NASA/ESA/Hubble Heritage

Part of a dataset collection on Hugging Face.

Dataset description

The Henry Draper Catalogue (HD) is the foundational reference for stellar spectral classification, containing 272,150 stars with spectral types assigned by Annie Jump Cannon at the Harvard College Observatory. Published in the Harvard Annals between 1918 and 1924, it established the OBAFGKM spectral sequence that remains in use today.

Annie Jump Cannon classified each star visually by examining the pattern of absorption lines on photographic plates, working at a rate of about 300 stars per hour. Her one-dimensional spectral sequence (O, B, A, F, G, K, M, plus R, N, S for carbon and zirconium-rich stars) encodes stellar surface temperature: O stars are the hottest (>30,000 K, ionised helium) and M stars the coolest (<3,500 K, molecular bands). The HD number (HD 1–272150) became the universal stellar identifier used in astrophysical literature throughout the 20th century and remains widely cited today.

Stars are classified to the nearest half subtype (e.g., G2V for the Sun). The luminosity class suffix — Ia (supergiant), II (bright giant), III (giant), IV (subgiant), V (main-sequence/dwarf) — was added for stars in the extension catalogs. The colour index B–V (photo_mag – photo_visual_mag) provides an independent temperature indicator correlated with spectral type, enabling photometric classification for very large samples.

This catalogue fills the key gap of stellar spectral types in the collection: no other dataset provides canonical MK classifications for a large, all-sky stellar sample. It is the natural complement to the Hipparcos parallax catalogue (astrometry), the Gaia DR3 datasets (photometry and radial velocities), and the GCVS variable star catalogue (variability). It enables HR-diagram studies, population synthesis, and training spectral-type classifiers.

This dataset is suitable for tabular classification tasks.

Schema

Column	Type	Description	Sample	Null %
hd_number	int64	Henry Draper Catalogue number (HD 1–272150) — the primary stellar identifier in this catalog	18667	0.0%
dm_designation	string	Durchmusterung identifier (BD, CD, or CP prefix) giving the Bonner/Córdoba/Cape catalog cross-match	BD-00 471	6.9%
photo_mag_quality	float64	Quality code for photo_mag: 0=normal, other codes indicate uncertain or combined measurements	0.0	14.4%
photo_mag	float64	Photographic magnitude (blue-sensitive plates, ~B band equivalent); brighter = lower number	8.7	14.3%
photo_visual_mag_quality	float64	Quality code for photo_visual_mag: 0=normal	1.0	3.9%
photo_visual_mag	float64	Photovisual magnitude (orthochromatic plates, ~V band equivalent)	9.7	0.5%
spectral_type	string	MK spectral type and luminosity class (e.g. G2V, K0III, A3m) — the primary classification product	K0	0.0%
intensity_code	string	Relative intensity indicator (1–6) used during classification; relates to plate exposure	4	19.4%
remarks	string	Remarks field: note on unusual classification, binary, or catalog flags	M	90.9%
ra_deg	float64	Right ascension (J2000, degrees) — computed by VizieR from B1900 coordinates	45.03110305555555	0.0%
dec_deg	float64	Declination (J2000, degrees) — computed by VizieR from B1900 coordinates	0.2479338888888888	0.0%
spectral_class	string	Broad spectral class (O/B/A/F/G/K/M) extracted from spectral_type; null for unusual types (W, R, N, S, C)	K	0.2%

Quick stats

• 272,150 stars with spectral type classifications (HD 1–272,150)
• 271,588 with OBAFGKM broad class; most common: A stars (72,155, 26.5%)
• Magnitudes span -1.6–50.0 (photographic)
• 272,150 stars with J2000 coordinates

Usage

from datasets import load_dataset

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

from datasets import load_dataset

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

# Spectral type distribution
import matplotlib.pyplot as plt
order = list("OBAFGKM")
counts = df["spectral_class"].value_counts().reindex(order, fill_value=0)
counts.plot(kind="bar", color="steelblue")
plt.xlabel("Spectral class")
plt.ylabel("Count")
plt.title("Henry Draper Catalogue — spectral class distribution")
plt.show()

# Colour-magnitude diagram (photo_mag vs B-V proxy)
import numpy as np
bv = df["photo_mag"] - df["photo_visual_mag"]
good = df[(df["photo_mag"] < 9) & bv.notna()]
plt.figure(figsize=(7, 8))
plt.scatter(bv[good.index], good["photo_mag"], s=0.3, alpha=0.2)
plt.gca().invert_yaxis()
plt.xlabel("B−V (photo_mag − photo_visual_mag)")
plt.ylabel("Photographic magnitude")
plt.title("HD Catalogue — colour-magnitude diagram")
plt.show()

Data source

https://vizier.cds.unistra.fr/viz-bin/VizieR-3?-source=III/135A/catalog

Related datasets

• juliensimon/hipparcos-catalog

• juliensimon/gaia-dr3-spectroscopic-binaries

• juliensimon/bright-stars

If you find this dataset useful, please consider giving it a like on Hugging Face. It helps others discover it.

About the author

Created by Julien Simon — AI Operating Partner at Fortino Capital. Part of the Space Datasets collection.

Citation

@dataset{henry_draper_catalog,
  title = {Henry Draper Catalogue},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/henry-draper-catalog},
  publisher = {Hugging Face}
}

License

CC-BY-4.0