SDOML-lite: SDO AIA and HMI Full-Disk Solar Images at 15-Minute Cadence, 2010-2024, for Machine Learning (Oxford)

Provider
oxai4science (University of Oxford), from NASA SDO AIA and HMI data
License
CC-BY-4.0 (source)
Size
5,199 files
Format
PNG (in sampled files)
Stored at
Hugging Face (https://huggingface.co/datasets/oxai4science/sdoml-lite)
Access
Free and public. No account with the data provider is needed.
Category
Astronomy & Space
Source
https://github.com/oxai4science/sdoml-lite

About this dataset

Full-disk images of the Sun from NASA's Solar Dynamics Observatory (SDO), resampled for machine learning in solar physics and space-weather forecasting. Each sample is one observation time at 15-minute cadence from 13 May 2010 to 31 July 2024 (about 500,000 timestamps) with six 512 x 512 float32 channels: extreme-ultraviolet and ultraviolet images from the Atmospheric Imaging Assembly (AIA) at 131, 171, 193, 211 and 1600 Angstrom, which trace hot coronal plasma, flares and the chromosphere, and a line-of-sight magnetogram, the surface magnetic field, from the Helioseismic and Magnetic Imager (HMI, hmi.M_720s). Inputs are AIA Level 1 FITS files and HMI data; AIA channels are normalized from statistics computed in a first processing pass. Files are WebDataset TAR archives, one per day (5,194 files, about 3 TB), each holding up to 96 observation sets as NumPy .npy files keyed YYYY/MM/DD/HHMM. SDOML-lite is a separate product from the SDOML dataset on AWS and is not format-compatible with it; a PyTorch loader is in the oxai4science/sdoml-lite GitHub repository. Uses: solar flare and coronal mass ejection forecasting, generative and self-supervised models of the Sun, active-region tracking, and EUV irradiance proxies. Released by oxai4science under CC BY 4.0; acknowledge NASA SDO and the AIA and HMI teams.

How to load this dataset in Python

from huggingface_hub import hf_hub_download
path = hf_hub_download(repo_id="oxai4science/sdoml-lite", filename="assets/ar1520.png", repo_type="dataset")
print(path)

Data structure

File types in sampled files: PNG

Files

Example file: assets/ar1520.png

Analyze this dataset without downloading it

On Scigantic, SDOML-lite: SDO AIA and HMI Full-Disk Solar Images at 15-Minute Cadence, 2010-2024, for Machine Learning (Oxford) opens in a hosted JupyterLab notebook with the data already mounted, so you can read files where they are stored instead of copying them first. Each notebook includes an AI assistant that can see the mounted files and write and run analysis code. Open the dataset page to launch a notebook, or browse the full catalog.

This page is also available as Markdown and as JSON.

Related Astronomy & Space datasets