Soil survey data for construction: the ground's own screening record, delivered

Three layers stacked inside one record, and each answers a different question about the same polygon.

The map unit layer is where everyone starts. One polygon per mapping area, keyed three ways: mukey is the unique key that joins everything else, muname is the human-readable name that encodes the slope range outright, and musym matches what a printed soil map shows.

The attribute stack beneath it is where the geotechnical detail lives: component-level shares and properties, then horizon-level attributes - the profile of each soil within the unit.

The interpretation layer on top turns all of it into engineering verdicts: suitability ratings for dwellings with basements, septic tank absorption fields and roads, resolved against a defined area of interest rather than eyeballed off a legend.

Datadory delivers all three from Web Soil Survey (WSS), USDA NRCS's authoritative soil survey record covering more than 95% of US counties. Inside our twenty-eight-record construction & engineering slice it scores 8 out of 10, and it is the only record that describes the ground itself - every other dataset in the slice measures something humans did to it.

Two real map units from the SSURGO database, exactly as they land in a delivery:

mukey: 49456
muname: Sarkar-McGilvery complex, 75 to 120 percent slopes
musym: 10F

mukey: 49457
muname: Kupreanof-Tolstoi association, 5 to 35 percent slopes
musym: 11B

One row per map unit polygon, and the two names above do real work before you open a single property column. "75 to 120 percent slopes" flags terrain you would not grade a building pad onto; "5 to 35 percent slopes" reads like ordinary hillside development. The name is a first-pass screen you get for nothing.

The complete dictionary ships with your sample, including the alternative area-of-interest keys - coordinates, a list of map unit keys, a soil survey area or an existing WSS area-of-interest id.

Geography is the headline. Soil maps span more than 95% of US counties plus US territories - roughly 100,000+ map unit polygons nationwide under one taxonomy. That single fact changes what a land team can attempt: a forty-parcel pipeline across four states resolves against one legend instead of four.

Granularity stacks three depths - polygon, component, horizon - along one join path. You walk from a parcel's map unit down to the soils that make up its share and their horizon properties without leaving the schema. Request ceilings run at 100,000 tabular rows or 250,000 spatial features, which is why long corridors arrive tiled into areas of interest rather than truncated mid-pull.

Five decisions separate a defensible screen from a casual lookup:

Name the geography and the asset type in a sample request and rows scoped to your parcels go out the same day.

A soil screen rarely stops at dirt, and three sibling records finish the picture inside the same industry slice.

Loads. USGS Earthquake Hazards API (FDSN) holds millions of worldwide event records reaching into the early 1900s for larger magnitudes, twenty-five documented fields apiece: origin time, epicenter, hypocenter depth, magnitude with its scale, quantified uncertainty, review status and PAGER alert level. Scored 9 out of 10, the highest in the construction & engineering slice.

Together the four records cover ground, loads, approvals and money - the four questions any site decision quietly asks.

Four records, four different questions - and choosing among them decides what your screen can claim:

Four limits belong in any methodology note, and all four favor teams that know them in advance.

A map unit is not a measurement at your footing line. It describes an association of soils across a polygon, and component shares vary inside it. Treat a favourable dwelling-with-basements rating as a screening signal that still needs borings and lab work to confirm - the granularity statement stops deliberately short of geotechnical truth.

Revisions are silent. Map units change where NRCS re-maps them, nothing announces it, and there is no release calendar to subscribe to. Snapshot discipline wins: capture retrieval dates with every extract and re-pull when revisions matter to the analysis.

Ratings are use-specific, not transferable. A septic tank absorption field verdict says nothing about subgrade, and a roads rating does not clear a basement. Match each requested interpretation to the decision it feeds instead of pulling one rating and generalizing.

Ceilings are real. Requests cap at 100,000 tabular rows or 250,000 spatial features, so corridors get tiled into areas of interest rather than discovered limits mid-pull. The same chunking discipline the seismic side imposes at its own request bounds.

Five personas get outsized value, and each works a different corner of the record.

Civil and geotechnical engineers use it the way nobody else can: as the first screen that decides whether a parcel deserves a drilling budget at all. Developers and data-product builders slot soil layers straight into property dashboards and due-diligence tools - the highest relevance fit in our catalog's construction slice. Data scientists and ML engineers train land-use and geotechnical models on a nationally consistent, mukey-keyed feature table with verified field definitions. Market researchers and consultants fold buildability context into feasibility reports without hand-tracing scanned soil maps. Journalists, academics and students report and teach from the authoritative national soil record itself.

Persona-by-persona workflow breakdowns live on developers builders use cases for construction engineering and data scientists use cases for construction engineering.

Because the hard part was never knowing SSURGO exists - it is keeping a polygon-plus-component-plus-horizon estate usable once it leaves the government app. Ratings arrive resolved to your area of interest instead of hand-read off legends; mukey joins hold steady between refreshes so a model trained this quarter still runs next quarter; and every extract ships with its retrieval date stamped, because a continuously revised database will not look identical next quarter.

Start with a sample: name the parcels, corridor or counties, and rows shaped exactly like the field dictionary go out sized to test in your own pipeline the same week.

Start with the construction engineering data hub, which indexes all twenty-eight records in the slice and shows where a ground-condition screen sits among them. Then go to the products themselves: Web Soil Survey (WSS) for the soil core and USGS Earthquake Hazards API (FDSN) for the loading history of the planet. The best construction & engineering datasets ranking scores the leaders side by side, the demand-side counterpart to a site screen is tracked monthly back to January 1990 in our Europe construction output index walkthrough, and the construction engineering data guide maps the wider pool - output statistics, permit records, safety enforcement and macro indicators.