nasa-apod / README.md
juliensimon's picture
Update NASA APOD: 11,268 entries
e4d8612 verified
metadata
license: cc-by-4.0
pretty_name: NASA Astronomy Picture of the Day
language:
  - en
description: >-
  Daily astronomy images and explanations from NASA's Astronomy Picture of the
  Day (APOD) program, the longest-running astronomy outreach initiative on the
  internet. Each entry features a curated astron
task_categories:
  - tabular-classification
  - text-classification
tags:
  - space
  - nasa
  - astronomy
  - apod
  - natural-language-processing
  - image
  - open-data
  - tabular-data
  - parquet
size_categories:
  - 10K<n<100K
configs:
  - config_name: default
    data_files:
      - split: train
        path: data/apod.parquet
    default: true

NASA Astronomy Picture of the Day

Blue Marble — high-definition image of Earth from space

Credit: NASA/GSFC/Suomi NPP

Part of a dataset collection on Hugging Face.

Dataset description

Daily astronomy images and explanations from NASA's Astronomy Picture of the Day (APOD) program, the longest-running astronomy outreach initiative on the internet. Each entry features a curated astronomical image or video with a written explanation by a professional astronomer.

APOD has published one entry every day since June 16, 1995 — an unbroken 30-year record spanning more than 10,000 entries. The subjects range from the mundane to the profound: sunsets, aurora, meteor showers, deep galaxy fields from the Hubble Space Telescope, gravitational wave detections, exoplanet atmosphere spectroscopy, and images from every NASA mission. Each explanation is written in accessible but scientifically accurate language, making the corpus particularly valuable for training and evaluating astronomy language models.

This dataset captures the full structured metadata for every APOD entry: date, title, explanation text, media URL, copyright status, and media type. Roughly 70% of entries feature NASA or ESA imagery in the public domain; the remaining ~30% are copyrighted contributions from amateur astrophotographers around the world. The url and hdurl fields provide stable links to the original images, enabling this metadata to be used as an index for downloading visual training data. The explanation field alone constitutes roughly 2 million words of expert-authored astronomy prose.

This dataset is suitable for tabular classification, text classification tasks.

Schema

Column Type Description Sample Null %
date str Date of the APOD entry in ISO format (YYYY-MM-DD); one entry per day since 1995-06-16; the primary key 1995-06-16 0.0%
title string Title of the astronomical image or media as chosen by the APOD editorial team; typically includes the object name or event described Neutron Star Earth 0.0%
explanation string Scientific explanation (~150-300 words) written by a professional astronomer; covers the physics, history, and significance of the featured subject; suitable for NLP, summarization, and astronomy education tasks Today's Picture: Explanation: If ... 0.1%
url string Primary media URL — direct image link or YouTube video URL; images are typically JPEG/PNG hosted on NASA GSFC servers https://apod.nasa.gov/apod/image/e_le... 0.2%
hdurl string Full-resolution (HD) image URL; null for video entries; typically 2000-8000px; may be very large https://apod.nasa.gov/apod/image/e_le... 3.7%
media_type string Content type: 'image' (majority) or 'video' (YouTube embeds); determines which URL fields are populated image 0.0%
copyright string Attribution string for the image author/photographer; null when the image is NASA, ESA, or other government agency public domain; ~30% of entries are copyrighted (amateur astrophotographers) 1989 by The University of Chicago. A... 50.7%
thumbnail_url string YouTube thumbnail URL for video entries; null for image entries; useful for visual preview without loading the full video https://img.youtube.com/vi/M6-iC_aYcu... 96.8%
is_public_domain bool True when copyright is null or empty, indicating NASA/ESA/public agency imagery with no usage restrictions; False for copyrighted astrophotography True 0.0%
year Int32 Calendar year of the entry (1995–present); derived from date; useful for time-series analysis and filtering by solar cycle or mission era 1995 0.0%
month Int32 Month (1–12) of the entry; derived from date; useful for seasonal sky analysis (e.g., summer Milky Way, winter Orion) 6 0.0%

Quick stats

  • 11,268 entries (1995-06-16 to 2026-04-26)
  • 10,847 images, 396 videos
  • 5,717 public domain entries (NASA/ESA), 5,551 copyrighted
  • Most common subjects in titles: nebula (1,029), galaxy (731), moon (627), mars (479), star (465)

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/nasa-apod", split="train")
df = ds.to_pandas()
from datasets import load_dataset
import pandas as pd

ds = load_dataset("juliensimon/nasa-apod", split="train")
df = ds.to_pandas()

# Public domain images only (safe to download/use)
public = df[df["is_public_domain"] == True]
print(f"{len(public):,} public domain entries with hdurl")

# Most popular topics by year
df["year"] = pd.to_numeric(df["year"])
topics = df[df["title"].str.contains("Galaxy|Nebula|Star|Comet|Moon", case=False, na=False)]
topics.groupby("year").size().plot(title="Cosmic Objects in APOD by Year")

# Recent entries
recent = df.sort_values("date").tail(10)[["date", "title", "media_type", "is_public_domain"]]
print(recent.to_string())

Data source

https://apod.nasa.gov/apod/

Update schedule

Daily at 07:00 UTC

Related datasets

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{nasa_apod,
  title = {NASA Astronomy Picture of the Day},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/nasa-apod},
  publisher = {Hugging Face}
}

License

CC-BY-4.0