Where the data comes from

Sources we use

OpenStreetMap (Overpass API)

Base coordinates, names, and occasional heights for every fall — 4,546 entries nationwide. OSM also supplies the polygon boundaries we use for National Park, National Recreation Area, and state-park containment checks.
License: ODbL. Attribution: © OpenStreetMap contributors.

USGS GNIS (Geographic Names Information System)

Canonical names and federal feature IDs from the United States Geological Survey. 1,392 falls carry a GNIS feature ID, used to anchor cross-references to other federal datasets.
License: public domain (U.S. Government work).

Wikidata + Wikipedia

Descriptions, official heights, history, and photo references. 624 falls are linked to a Wikidata entity; 425 carry a Wikipedia description.
License: CC0 for Wikidata facts; CC BY-SA for Wikipedia text — we attribute the source article on each fall page.

Wikimedia Commons

Photos with photographer credit and license. 417 falls carry a Commons photo, each rendered with author + license + link back to the source page.
License: per-image (typically CC BY, CC BY-SA, or public domain). Hover the credit line on any photo for the specific license.

National Park Service

Park membership, entrance fees, and live alerts for falls inside NPS units. 509 falls sit inside a national park, monument, seashore, or recreation area; 477 carry a current NPS alert (refreshed monthly).
License: public domain (U.S. Government work).

State-park boundaries

State-park containment is computed from OSM polygons (we filter relations whose name contains "State Park" or "State Recreation Area" — the only tag that's reliably consistent across all 50 states). 299 falls sit inside a tagged state park. Coverage varies by how completely each state's parks have been mapped in OpenStreetMap.

What we deliberately don't do

• We don't fabricate safety information. Swim policies, dog rules, closures, and seasonal access are shown only when sourced from an authoritative feed (NPS alerts, etc). Anything we don't know stays blank.
• We don't scrape copyrighted blog or business photos. Photos come only from Wikimedia Commons with the photographer credited and license preserved.
• We don't claim a fall is in a park unless a polygon says so. Containment is a point-in-polygon test, never a guess based on proximity to a park's centroid.

Refresh cadence

The full pipeline (OSM base pull + GNIS, Wikidata, NPS, state-park enrichment) runs on the first of each month via GitHub Actions. The pull-request diff is reviewed by hand before the dataset goes live.

Something off?

Spotted a wrong location, missing fall, or stale info? Tell us here. Most factual fixes belong upstream — at OpenStreetMap or the relevant Wikidata/Wikipedia article — and will flow into our next monthly refresh.