juliensimon/donki-space-weather-events

NASA DONKI Space Weather Events

Credit: NASA

Part of a dataset collection on Hugging Face.

Dataset description

Space weather events from NASA's DONKI (Database Of Notifications, Knowledge, Information) at the Community Coordinated Modeling Center. Covers coronal mass ejections, geomagnetic storms, interplanetary shocks, high-speed streams, and solar energetic particles from 2010 to present.

DONKI tracks the chain of space weather events from Sun to Earth. A CME erupts from the solar corona at speeds ranging from 250 to over 3,000 km/s, driving an interplanetary shock (IPS) ahead of it. When the shock and CME arrive at Earth, they compress the magnetosphere and produce a geomagnetic storm (GST) measurable via the Kp and Dst indices. High-speed streams (HSS) from coronal holes produce recurring disturbances on a ~27-day cadence, while solar energetic particle (SEP) events deliver MeV-range protons within minutes to hours of the initiating flare or CME.

DONKI is uniquely valuable because it preserves the causal linkages between these phenomena via the linked_events field. Unlike raw index time series, DONKI records which specific CME triggered which geomagnetic storm, making it possible to study transit times, geoeffectiveness as a function of CME speed and direction, and the statistical reliability of CME arrival forecasts.

This dataset is suitable for tabular classification, time-series forecasting tasks.

Schema

Column	Type	Description	Sample	Null %
event_type	str	Event category: CME (coronal mass ejection), GST (geomagnetic storm), IPS (interplanetary shock), HSS (high-speed stream), or SEP (solar energetic particle)	IPS	0.0%
activity_id	str	Unique DONKI event identifier (e.g. '2024-05-08T22:09:00-CME-001'); primary key for cross-referencing	2010-01-20T20:20:00-IPS-001	0.0%
start_time	datetime64[ns, UTC]	Event start time in UTC; for CMEs this is the first coronagraph appearance, for GSTs the storm onset	2010-01-20 20:20:00+00:00	0.0%
source_location	str	Solar source location in heliographic coordinates (e.g. 'N23W45'); CME-only, null for other event types	S20E05	73.6%
active_region	Int64	NOAA active region number associated with the event (e.g. 13664); CME-only, null for other types	11123	84.8%
note	str	Analyst notes from CCMC space weather forecasters; may contain event details or IPS location info	STEREO B	11.7%
link	str	URL to the DONKI web page for this specific event; useful for accessing additional details and linked analyses	https://kauai.ccmc.gsfc.nasa.gov/DONK...	0.0%
cme_speed_kms	float64	CME speed in km/s from coronagraph analysis (SOHO/LASCO or STEREO/COR); ranges from ~250 to >3000 km/s; CME-only	620.0	23.2%
cme_half_angle_deg	float64	CME angular half-width in degrees from coronagraph imagery; halo CMEs have half-angle near 90 degrees; CME-only	26.0	23.2%
cme_latitude	float64	CME source latitude in degrees from coronagraph analysis; CME-only	7.0	23.3%
cme_longitude	float64	CME source longitude in degrees from coronagraph analysis; CME-only	8.0	33.7%
cme_type	str	CME morphological type: S (slow), C (common), O (occasional), R (rare), ER (extremely rare); CME-only	C	23.2%
cme_time_21_5	datetime64[ns, UTC]	Estimated time the CME reaches 21.5 solar radii (roughly 0.1 AU); used for transit-time modeling; CME-only	2010-04-03 17:16:00+00:00	23.2%
cme_measurement	str	Coronagraph measurement technique used for CME parameter estimation; CME-only	null	23.2%
linked_events	str	Comma-separated activity IDs of causally linked events; enables Sun-to-Earth chain analysis (e.g. CME -> IPS -> GST)	2010-04-03T09:04:00-FLR-001, 2010-04-...	68.6%
gst_max_kp	float64	Maximum Kp index recorded during the geomagnetic storm (0-9 scale); Kp >= 5 is a minor storm, >= 7 strong, 9 extreme; GST-only	7.0	98.4%
gst_kp_count	float64	Number of 3-hour Kp index readings during the storm duration; GST-only	1.0	98.4%

Quick stats

• 12,230 events (2010-01-20 to 2026-04-18)
• 9,464 CMEs, 194 geomagnetic storms, 1,340 interplanetary shocks
• 764 high speed streams, 468 solar energetic particle events
• Fastest CME: 3529 km/s on 2024-02-12

Usage

from datasets import load_dataset

ds = load_dataset("juliensimon/donki-space-weather-events", split="train")
df = ds.to_pandas()

from datasets import load_dataset

ds = load_dataset("juliensimon/donki-space-weather-events", split="train")
df = ds.to_pandas()

# Fast CMEs (potential Earth-directed storms)
fast_cmes = df[(df["event_type"] == "CME") & (df["cme_speed_kms"] > 1000)]

# Geomagnetic storms with linked CMEs
storms = df[df["event_type"] == "GST"]
storms_with_cme = storms[storms["linked_events"].str.contains("CME", na=False)]

# CME speed distribution
import matplotlib.pyplot as plt
cmes = df[df["event_type"] == "CME"]
cmes["cme_speed_kms"].hist(bins=50)
plt.xlabel("CME Speed (km/s)")
plt.ylabel("Count")
plt.title("DONKI CME Speed Distribution")
plt.show()

# Event frequency by type and year
df["year"] = df["start_time"].dt.year
df.groupby(["year", "event_type"]).size().unstack().plot()
plt.title("DONKI Events by Type and Year")
plt.show()

Data source

https://ccmc.gsfc.nasa.gov/tools/DONKI/

Update schedule

Daily at 14:00 UTC via GitHub Actions

Related datasets

• juliensimon/solar-flare-events

• juliensimon/space-weather-indices

• juliensimon/dst-index

• juliensimon/neo-close-approaches

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{donki_space_weather_events,
  title = {NASA DONKI Space Weather Events},
  author = {juliensimon},
  year = {2026},
  url = {https://huggingface.co/datasets/juliensimon/donki-space-weather-events},
  publisher = {Hugging Face}
}

License

CC-BY-4.0