Hugging Face's logo

Datasets:
lukeslp
/
us-disasters-mashup

Tasks:
Feature Extraction
Time Series Forecasting
Modalities:
Geospatial
Size:
10K<n<100K
Tags:
disasters
aviation
maritime
weather
seismic
geospatial
License:
Dataset card Files
xet
 Community
Dataset Viewer
Duplicate
The dataset viewer is not available for this split.
Cannot load the dataset split (in streaming mode) to extract the first rows.
Error code:   StreamingRowsError
Exception:    CastError
Message:      Couldn't cast
dataset_name: string
description: string
last_updated: timestamp[s]
source: string
source_url: string
license: string
record_count: int64
file_format: string
file_size_mb: double
date_range: string
fields: struct<Date: string, Time: string, Location: string, Operator: string, Flight #: string, Route: stri (... 123 chars omitted)
  child 0, Date: string
  child 1, Time: string
  child 2, Location: string
  child 3, Operator: string
  child 4, Flight #: string
  child 5, Route: string
  child 6, Type: string
  child 7, Registration: string
  child 8, cn/In: string
  child 9, Aboard: string
  child 10, Fatalities: string
  child 11, Ground: string
  child 12, Summary: string
notable_records: struct<first_fatality: string, largest_single_aircraft: string>
  child 0, first_fatality: string
  child 1, largest_single_aircraft: string
visualization_suggestions: list<item: string>
  child 0, item: string
category: null
latitude: null
longitude: null
name: null
date: null
subcategory: null
magnitude: null
fatalities: null
injuries: null
damage: null
state: null
aircraft_type: null
event_id: null
vessel_type: null
depth_km: null
to
{'category': Value('string'), 'latitude': Value('float64'), 'longitude': Value('float64'), 'name': Value('string'), 'date': Value('string'), 'subcategory': Value('string'), 'magnitude': Value('float64'), 'fatalities': Value('int64'), 'injuries': Value('int64'), 'damage': Value('string'), 'state': Value('string'), 'aircraft_type': Value('string'), 'event_id': Value('string'), 'vessel_type': Value('string'), 'depth_km': Value('float64')}
because column names don't match
Traceback:    Traceback (most recent call last):
                File "/src/services/worker/src/worker/utils.py", line 99, in get_rows_or_raise
                  return get_rows(
                         ^^^^^^^^^
                File "/src/libs/libcommon/src/libcommon/utils.py", line 272, in decorator
                  return func(*args, **kwargs)
                         ^^^^^^^^^^^^^^^^^^^^^
                File "/src/services/worker/src/worker/utils.py", line 77, in get_rows
                  rows_plus_one = list(itertools.islice(ds, rows_max_number + 1))
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2543, in __iter__
                  for key, example in ex_iterable:
                                      ^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2060, in __iter__
                  for key, pa_table in self._iter_arrow():
                                       ^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/iterable_dataset.py", line 2092, in _iter_arrow
                  pa_table = cast_table_to_features(pa_table, self.features)
                             ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
                File "/usr/local/lib/python3.12/site-packages/datasets/table.py", line 2192, in cast_table_to_features
                  raise CastError(
              datasets.table.CastError: Couldn't cast
              dataset_name: string
              description: string
              last_updated: timestamp[s]
              source: string
              source_url: string
              license: string
              record_count: int64
              file_format: string
              file_size_mb: double
              date_range: string
              fields: struct<Date: string, Time: string, Location: string, Operator: string, Flight #: string, Route: stri (... 123 chars omitted)
                child 0, Date: string
                child 1, Time: string
                child 2, Location: string
                child 3, Operator: string
                child 4, Flight #: string
                child 5, Route: string
                child 6, Type: string
                child 7, Registration: string
                child 8, cn/In: string
                child 9, Aboard: string
                child 10, Fatalities: string
                child 11, Ground: string
                child 12, Summary: string
              notable_records: struct<first_fatality: string, largest_single_aircraft: string>
                child 0, first_fatality: string
                child 1, largest_single_aircraft: string
              visualization_suggestions: list<item: string>
                child 0, item: string
              category: null
              latitude: null
              longitude: null
              name: null
              date: null
              subcategory: null
              magnitude: null
              fatalities: null
              injuries: null
              damage: null
              state: null
              aircraft_type: null
              event_id: null
              vessel_type: null
              depth_km: null
              to
              {'category': Value('string'), 'latitude': Value('float64'), 'longitude': Value('float64'), 'name': Value('string'), 'date': Value('string'), 'subcategory': Value('string'), 'magnitude': Value('float64'), 'fatalities': Value('int64'), 'injuries': Value('int64'), 'damage': Value('string'), 'state': Value('string'), 'aircraft_type': Value('string'), 'event_id': Value('string'), 'vessel_type': Value('string'), 'depth_km': Value('float64')}
              because column names don't match

Need help to make the dataset viewer work? Make sure to review how to configure the dataset viewer, and open a discussion for direct support.

US Disasters Mashup

54,575 disaster events pulled from four US government databases into one flat JSON file. Plane crashes, shipwrecks, tornadoes, earthquakes -- all geocoded and categorized.

What's In It

Category Records Source Date Range
Aviation Accidents 32,410 NTSB 1974-2018
Severe Storms 14,770 NOAA Storm Events 1950-2025
Earthquakes 3,742 USGS Ongoing
Shipwrecks 3,653 NOAA AWOIS Historical (1600s-1970s)

Fields

Every record has core fields (category, latitude, longitude, name, date, subcategory). Additional fields depend on the category:

Field Coverage Which Categories
category 100% All
latitude / longitude 100% All
name 100% All -- event description or location
subcategory 100% Tornado, Flash Flood, seismic, maritime, aviation, etc.
date 94% All except some historical shipwrecks
aircraft_type 59% Aviation only
event_id 59% Aviation only (NTSB event IDs)
magnitude 20% Storms (Fujita/EF scale) + Earthquakes (Richter)
fatalities 27% Storms
injuries 27% Storms
damage 26% Storms (text format: "250K", "1.5M")
state 27% Storms
vessel_type <1% Shipwrecks (sparse)

Example Records

Storm:

{
  "category": "storm",
  "latitude": 34.88,
  "longitude": -99.28,
  "name": "Tornado in OKLAHOMA, KIOWA",
  "date": "1950-04-28",
  "subcategory": "Tornado",
  "magnitude": "0",
  "fatalities": "1",
  "injuries": "1",
  "damage": "250K",
  "state": "OKLAHOMA"
}

Aviation:

{
  "category": "aviation_accident",
  "latitude": 20.000833,
  "longitude": -155.6675,
  "name": "Aviation Accident - SCHLEICHER ASH25M",
  "date": "2012-01-01",
  "subcategory": "aviation",
  "aircraft_type": "SCHLEICHER ASH25M",
  "event_id": "20121010X84549"
}

Known Quirks

A few things worth knowing if you're working with this data:

  • Aviation dates are year-only. All 32,410 aviation records show YYYY-01-01. The actual dates are embedded in the event IDs (e.g., 20121010X84549 = Oct 10, 2012) but the date field just has the year.
  • Earthquake dates are Unix timestamps, not ISO format. Convert with datetime.fromtimestamp().
  • ~5,983 duplicate aviation records from overlapping source files. Deduplicate on event_id if you need unique events.
  • Coordinates extend beyond CONUS. About 3,200 records are in Hawaii, Alaska, territories, or international waters. Expected for aviation and maritime data.
  • depth_km is always null. The field exists in the schema but was never populated.

Loading 

import json

with open("disasters_mashup.json") as f:
    disasters = json.load(f)

# Filter by category
storms = [d for d in disasters if d["category"] == "storm"]
aviation = [d for d in disasters if d["category"] == "aviation_accident"]

# Deduplicate aviation (optional)
seen = set()
unique_aviation = []
for d in aviation:
    if d.get("event_id") not in seen:
        seen.add(d.get("event_id"))
        unique_aviation.append(d)

Sources

All public domain, from US government agencies:

Where to Get It

License

CC0 1.0 (Public Domain). All source data comes from US government agencies.

Author

Luke Steuber

Citation 

@dataset{steuber2026disasters,
  title={US Disasters Mashup},
  author={Steuber, Luke},
  year={2026},
  publisher={GitHub/HuggingFace/Kaggle},
  url={https://github.com/lukeslp/us-disasters-mashup}
}

Structured Data (JSON-LD) 

{
  "@context": "https://schema.org",
  "@type": "Dataset",
  "name": "US Disasters Mashup",
  "description": "54,575 disaster events from four US government databases (NTSB aviation accidents, NOAA shipwrecks, NOAA severe storms, USGS earthquakes) unified into a single geocoded JSON file.",
  "url": "https://github.com/lukeslp/us-disasters-mashup",
  "sameAs": [
    "https://huggingface.co/datasets/lukeslp/us-disasters-mashup",
    "https://www.kaggle.com/datasets/lucassteuber/us-disasters-mashup"
  ],
  "license": "https://creativecommons.org/publicdomain/zero/1.0/",
  "creator": {
    "@type": "Person",
    "name": "Luke Steuber",
    "url": "https://lukesteuber.com"
  },
  "keywords": ["disasters", "aviation accidents", "shipwrecks", "storms", "earthquakes", "geospatial", "united states"],
  "temporalCoverage": "1600/2025",
  "spatialCoverage": {
    "@type": "Place",
    "name": "United States"
  },
  "distribution": [
    {
      "@type": "DataDownload",
      "encodingFormat": "application/json",
      "contentUrl": "https://github.com/lukeslp/us-disasters-mashup"
    }
  ]
}
Downloads last month
30