Reference Reference tableQC clean AI-training-safe

US Storm Events

The official U.S. record of severe weather, cleaned into one fat events table: tornadoes, hail, floods, hurricanes, winter storms, heat, wildfire - every event since 2000 with the human and financial toll. NOAA/NCEI ships the Storm Events Database as one gzipped ~50-column CSV per year, with property and crop damage encoded as strings like '1.50M'; we assemble every year, parse the damages to real USD numbers, and keep the fields that matter - timing, state, hazard type, injuries and deaths (direct & indirect), property & crop damage in USD, magnitude, tornado F/EF scale, and a begin geocode. One row per event, keyed by NOAA event id. The open, authoritative alternative to CC-BY-NC EM-DAT. Public domain.

1.6M
rows
21
columns
64.8 MB
download
Source
NOAA
Licence
Redistributable open data
Quality
QC clean · 0 failures
Format
Apache Parquet + dictionary

Safe to train on

Built only from a named, redistributable official source under a documented open licence - not scraped web data. No copyright grey zone, no personal data. Ships machine-readable Croissant metadata (ML Commons - loads in Hugging Face / Kaggle / Google), an AI training-licence manifest documenting source, licence and provenance for your model's data governance, and a machine-readable data dictionary (drop it into an agent / RAG prompt so the model knows every column) - all inspectable before you buy.

What's included

  • Every U.S. severe-weather event since 2000 - one fat, tidy row per event (~1.6M events)
  • Injuries & deaths (direct + indirect); property & crop damage PARSED to real USD
  • Hazard type, magnitude, tornado F/EF scale, county/zone, WFO office, begin lat/lon
  • Damage strings like '1.50M' cleaned to numeric USD - the tedious part, done
  • Audiences: (re)insurance & cat-risk, climate-risk, agriculture, resilience/gov
  • The open, authoritative alternative to CC-BY-NC-ND EM-DAT
  • Public domain (U.S. NOAA/NCEI) - free commercial reuse & resale
  • Source: NOAA National Centers for Environmental Information - Storm Events Database

Schema

ColumnTypeDescriptionFilledDistinct
event_id int NOAA Storm Events event id (primary key) 100.0% 1,610,092
episode_id int NOAA episode id (groups related events) 100.0% 318,068
begin_datetime str Event begin time (YYYY-MM-DD HH:MM:SS) 100.0% 821,364
end_datetime str Event end time (YYYY-MM-DD HH:MM:SS) 100.0% 803,876
state str U.S. state (uppercase) 100.0% 69
year int Year of the event 100.0% 27
event_type str Hazard type (Tornado, Hail, Flash Flood, Winter Storm, …) 100.0% 56
cz_name str County/zone name 100.0% 5,373
wfo str NWS Weather Forecast Office id 100.0% 125
injuries_direct int Direct injuries 100.0% 135
injuries_indirect int Indirect injuries 100.0% 66
deaths_direct int Direct deaths 100.0% 48
deaths_indirect int Indirect deaths 100.0% 22
damage_property_usd float Property damage, parsed to USD 69.9% 2,184
damage_crops_usd float Crop damage, parsed to USD 65.2% 1,123
magnitude float Event magnitude (hail size in inches / wind speed in knots) 52.6% 446
magnitude_type str Magnitude type (EG, MG, ES, MS) 34.1% 6
tor_f_scale str Tornado F/EF scale (where applicable) 2.3% 12
begin_lat float Begin latitude 59.5% 134,564
begin_lon float Begin longitude 59.5% 189,132
source str Report source (e.g. Trained Spotter, Public, Emergency Manager) 100.0% 71

Sample & preview

Every purchase ships as Apache Parquet with a data dictionary and the full QC report. A free sample (first rows + schema) is downloadable here - confirm fit before you buy. Source: NOAA (redistributable open data; attribution passes through - see our Licence).

Don't trust screenshots - drop the sample into your notebook right now.

import pandas as pd
df = pd.read_parquet("us-storm-events_sample.parquet")
df.info()   # typed columns, gap-aware, ready to join
⬇ Download free Parquet sample or CSV for Excel