CoolFace
Datasetpublic

sachinkg12/us-county-hazard-features

US County Multi-Hazard Features Dataset A curated, ML-ready dataset of 1,014,930 county-month observations spanning 3,222 US counties from 2000 to 2026, integrating 7 federal data sources into 42 engineered features for predicting FEMA disaster declarations 90 days in advance. Dataset Summary Property Value Rows 1,014,930 Columns 50 (42 features + target + metadata) Counties 3,222 (all US counties with available data) Time span 2000-01 to 2026-03… See the full description on the dataset page: https://huggingface.co/datasets/sachinkg12/us-county-hazard-features.

sourceHugging Faceapache-2.0updated 6mo agoView on Hugging Face
0likes16downloads
Dataset Card

US County Multi-Hazard Features Dataset

A curated, ML-ready dataset of 1,014,930 county-month observations spanning 3,222 US counties from 2000 to 2026, integrating 7 federal data sources into 42 engineered features for predicting FEMA disaster declarations 90 days in advance.

Dataset Summary

PropertyValue
Rows1,014,930
Columns50 (42 features + target + metadata)
Counties3,222 (all US counties with available data)
Time span2000-01 to 2026-03
GranularityCounty-month
Targetdeclaration_next_90d — binary, FEMA disaster declaration within 90 days
Positive rate11.02%
FormatApache Parquet (Snappy compression)

Data Sources

This dataset integrates 7 federal data sources — all free, no API keys required:

SourceWhat
FEMA Disaster DeclarationsEvery federal disaster declaration by county (1953–present)
USGS EarthquakesM2.5+ seismic events with coordinates (1964–present)
NOAA Storm EventsTornadoes, floods, hurricanes, hail + casualties/damage (2000–present)
US Census BureauCounty demographics, housing, economics
US Drought MonitorWeekly drought severity by county, D0–D4 (2000–present)
NIFC WildfiresWildfire incidents with acres burned (2000–present)
NFIP Flood ClaimsNational Flood Insurance Program claims and payouts (1978–present)

Features (42 total)

FEMA History (7)

Rolling window declaration counts and recency metrics.

FeatureTypeDescription
declarations_1yrintDisaster declarations in prior 1 year
declarations_3yrintDisaster declarations in prior 3 years
declarations_5yrintDisaster declarations in prior 5 years
declarations_10yrintDisaster declarations in prior 10 years
months_since_last_declintMonths since most recent declaration (-1 if none)
major_disaster_ratiofloatFraction of declarations that were major disasters
ia_program_ratiofloatFraction with Individual Assistance programs

Storm Events (10)

NOAA severe weather aggregations.

FeatureTypeDescription
storm_event_count_1yrintStorm events in prior 1 year
storm_event_count_5yrintStorm events in prior 5 years
storm_deaths_5yrintStorm-related deaths in prior 5 years
storm_injuries_5yrintStorm-related injuries in prior 5 years
storm_property_damage_5yrfloatProperty damage ($) in prior 5 years
storm_crop_damage_5yrfloatCrop damage ($) in prior 5 years
tornado_count_5yrintTornado events in prior 5 years
flood_count_5yrintFlood events in prior 5 years
hail_count_5yrintHail events in prior 5 years
max_tor_f_scale_5yrintMaximum tornado F-scale in prior 5 years

Socioeconomic (5)

US Census demographic and economic indicators.

FeatureTypeDescription
populationlongCounty population
housing_unitslongNumber of housing units
median_home_valuelongMedian home value ($)
population_densityfloatPeople per square mile
land_area_sq_mifloatCounty land area in square miles

Drought (4)

US Drought Monitor severity metrics.

FeatureTypeDescription
drought_severity_avg_5yrfloatAverage drought severity score (5yr)
drought_max_severity_5yrfloatMaximum drought severity score (5yr)
severe_drought_weeks_5yrintWeeks of severe drought (D2+) in 5 years
drought_d4_pct_max_5yrfloatPeak percentage of county in D4 (exceptional) drought

Wildfire (4)

NIFC wildfire incident metrics.

FeatureTypeDescription
wildfire_count_1yrintWildfire incidents in prior 1 year
wildfire_count_5yrintWildfire incidents in prior 5 years
wildfire_acres_burned_5yrfloatTotal acres burned in prior 5 years
wildfire_max_acres_5yrfloatLargest single wildfire (acres) in 5 years

NFIP Flood Insurance (3)

National Flood Insurance Program claim patterns.

FeatureTypeDescription
nfip_claim_count_5yrintNFIP claims in prior 5 years
nfip_total_payout_5yrfloatTotal NFIP payouts ($) in prior 5 years
nfip_avg_payout_5yrfloatAverage NFIP payout ($) in prior 5 years

Spatial (2)

Neighborhood and state-level context.

FeatureTypeDescription
neighbor_avg_declarations_5yrfloatAverage 5yr declarations of neighboring counties
state_avg_declarations_5yrfloatAverage 5yr declarations across the state

Cascade Interaction Features (7)

Novel contribution: Multiplicative interaction terms capturing multi-hazard co-occurrence.

FeatureTypeDescription
cascade_drought_fire_riskfloatdrought_severity_6mo × log1p(wildfire_acres_1yr)
cascade_fire_flood_riskfloatlog1p(burn_scar_acres_18mo) × flood_events_1yr
cascade_hurricane_flood_riskfloathurricane_declarations_60d × flood_events_30d
cascade_earthquake_landslide_riskfloatsignificant_quakes_90d × severe_storms_30d
cascade_storm_compound_countintSevere storms in prior 30 days (compound events)
cascade_active_chainsintCount of active cascade interactions (0–5)
cascade_max_chain_lengthintLongest active hazard chain (1–3)

Target & Metadata

ColumnTypeDescription
fipsstring5-digit FIPS county code
year_monthstringObservation month (YYYY-MM)
declaration_next_90dboolTarget: FEMA declaration within 90 days
declaration_type_next_90dstringDeclaration type if positive (DR, EM, etc.)

Usage

python
import pandas as pd

df = pd.read_parquet("us-county-hazard-features.parquet")

# Temporal train/test split (recommended)
train = df[df["year_month"] < "2022-01"]
val = df[(df["year_month"] >= "2022-01") & (df["year_month"] < "2023-01")]
test = df[(df["year_month"] >= "2023-01") & (df["year_month"] <= "2024-12")]
# Note: Exclude months after 2024-12 — FEMA declaration data is incomplete

# Feature columns (42 features, no temporal — see paper for ablation justification)
FEATURE_COLS = [c for c in df.columns if c not in [
    "fips", "year_month", "declaration_next_90d", "declaration_type_next_90d",
    "month_of_year", "is_hurricane_season", "is_tornado_season", "is_wildfire_season"
]]

Benchmark Results

Using XGBoost with temporal split:

ModelROC-AUCPR-AUCF1
Naive (prior)0.5000.5490.000
Logistic Regression0.5420.1390.199
Random Forest0.8450.2870.334
XGBoost0.8930.5550.482

95% Bootstrap CI: ROC-AUC [0.890, 0.896]

Key Findings

  1. 1.FEMA Dominance: Removing FEMA features drops AUC from 0.89 to 0.63 — declaration history is the strongest predictor, suggesting the federal process is path-dependent.
  1. 1.Cascade Interactions: Multi-hazard cascade features improve compound disaster detection (recall lift +2.9% for cascade events, ROC-AUC 0.907 vs 0.893 overall).
  1. 1.Declaration Equity: Low-income counties (Q1) show 2.3x higher prediction residuals than wealthy counties (Q4) at the same hazard exposure level (p < 1e-100), suggesting structural inequities in federal disaster declarations.

Citation

bibtex
@dataset{gupta2026uscountyhazard,
  title={US County Multi-Hazard Features for Disaster Declaration Prediction},
  author={Gupta, Sachin},
  year={2026},
  publisher={Hugging Face},
  url={https://huggingface.co/datasets/sachinkg12/us-county-hazard-features}
}

License

Apache 2.0

Source Code

HazardCast — Full pipeline: data ingestion, feature engineering, model training, and REST API.