Datasetpublic

sweetapricity/bridgedeck-statsgo

Bridgedeck — STATSGO Soil Properties Soil properties at ~906k US bridge locations, queried from the NRCS Soil Data Access (SDA) REST API using the STATSGO2/SSURGO national database. Source Database: NRCS Soil Data Access (SDA) — https://SDMDataAccess.sc.egov.usda.gov Spatial data: STATSGO2 / SSURGO national map units Query method: Point-in-polygon via SDA_Get_Mukey_from_intersection_with_WktWgs84 Coverage: All unique NBI bridge locations in CONUS (~906k points)… See the full description on the dataset page: https://huggingface.co/datasets/sweetapricity/bridgedeck-statsgo.

sourceHugging Faceotherupdated 4mo agoView on Hugging Face

0likes11downloads

Dataset Card

Bridgedeck — STATSGO Soil Properties

Soil properties at ~906k US bridge locations, queried from the NRCS Soil Data Access (SDA) REST API using the STATSGO2/SSURGO national database.

Source

—Database: NRCS Soil Data Access (SDA) — https://SDMDataAccess.sc.egov.usda.gov
—Spatial data: STATSGO2 / SSURGO national map units
—Query method: Point-in-polygon via SDA_Get_Mukey_from_intersection_with_WktWgs84
—Coverage: All unique NBI bridge locations in CONUS (~906k points)

Schema

Column	Type	Description
`bridge_id`	str	Bridge identifier (`{state}_{county}_{structure}`)
`latitude`	float	Bridge latitude (WGS-84)
`longitude`	float	Bridge longitude (WGS-84)
`mukey`	str	STATSGO/SSURGO map unit key
`compname`	str	Dominant soil component name
`comppct_r`	int	Component percentage of map unit
`corsteel`	str	Corrosion potential for steel (High/Moderate/Low)
`corcon`	str	Corrosion potential for concrete (High/Moderate/Low)
`drainagecl`	str	Drainage class (Well drained / Somewhat poorly drained / etc.)
`hydgrp`	str	Hydrologic group (A/B/C/D) — relates to water retention
`taxorder`	str	Soil taxonomy order (Mollisols, Alfisols, etc.)
`silttotal_r`	float	Surface horizon silt percentage (%) — >50% indicates frost susceptibility

Derived Features

The ETL script (statsgo_clean.py) derives:

—soil_type: Frost susceptibility classification based on silt content and hydrologic group

Notes

—Urban bridges (~30%) return empty soil data ("Urban land" has no soil properties)
—These receive NaN in the derived features — XGBoost handles NaN natively
—Data is static (soil properties don't change year to year)

Related Datasets

—bridgedeck-nbi — bridge locations used for spatial query
—bridgedeck-elevation — complementary physical property

Citation

USDA-NRCS. Soil Survey Staff. Web Soil Survey. https://websoilsurvey.sc.egov.usda.gov/