Filters parcels & buildings (z10+); county/hex aggregates are unaffected.
Zoom in to drill down: counties → H3 hexes → parcels → individual buildings. Aggregates shade by the selected peril's elevated share, so they move with the toggle. Click a county to focus the panel on it.
Read before quoting.
· yearbuilt is construction year, not roof age.
· Hail is radar-detected density, not modelled severity.
· Tornado / wind / winter are tract-resolution, not per-building.
· Hazard is centroid-sampled for 98.24% of buildings.
· Assessor $ values are popup-only reference (≈63% coverage, county-dependent); never used for styling or ranking.
· NULL road/station distance = furthest (no station within 25 km), not missing —
on a building, and on a parcel that has buildings. A parcel with no buildings
reads “no building to measure”: nothing was measured there, which is not the
same claim as remoteness.
· Access drive time is a modelled estimate in 5-minute bands — straight-line
distance × a terrain factor calibrated against HERE routes, not a per-building route.
· Aerial imagery is locational context only, including at partial opacity —
blending it under the footprints changes what you can SEE, never what was measured. Scores are screening indicators from
modelled and tract-resolution inputs — never derived from imagery. A building shown
under aerial is not individually assessed. The aerial is NAIP (a mosaic of flights from 2022 on,
so the image date varies by area) with USGS National Map imagery alone when zoomed out and
under NAIP while it loads.
Statewide exposure
Counties by exposure
Assessor $ coverage — why not dollars
Improvement-value coverage is ~63% and county-dependent. The largest exposures often publish the least — El Paso (the state's biggest, 290,630 buildings) reports $0. A dollar ramp maps who reports, not where risk is.