Sources and methodology
Sources, in order of authority
- Official state park sources. Each state’s park directory, park pages, alerts pages and official data services (some states publish JSON or ArcGIS data). These decide which parks exist and supply hours, fees, rules, amenities, camping and alerts.
- Official open data from state GIS portals, used for inventory and coordinates where a state’s website blocks automated access or lacks them.
- USGS PAD-US 4.1 (public domain) for park boundaries and approximate acreage.
- Wikidata (CC0) for cross-checking the inventory, alternate names, coordinates and establishment dates, and for finding photos.
- US Census Bureau gazetteer and boundary files for towns, ZIP codes and counties.
When sources disagree, the official state source wins. Secondary sources only fill gaps, and the page says which source a fact came from (for example, coordinates “from Wikidata”). We never publish a fact found only in unofficial reviews or crowd-sourced listings.
How often information is checked
| Information | Checked |
|---|---|
| Statewide alerts and closure feeds | Every 2 hours |
| Park pages for parks with active alerts | Daily |
| All other park pages (hours, fees, reservations, pets, amenities, camping) | At least every 3 days, in rotation |
| Official inventory (new, renamed, reclassified or closed parks) | Weekly |
| Statewide pass and reservation pages | Weekly (changes are flagged for review) |
| Public-domain status of every photo in use (removed automatically if it changes) | Monthly |
| Searching for new public-domain photos | In periodic batches |
Where a state’s site asks crawlers to slow down (a robots.txt crawl delay), we follow it, so a few states’ park pages refresh more slowly than this schedule — New York’s, for example, take about two days per full pass.
“Last checked” vs. “last changed.” “Last checked” is when we last read the source successfully. “Last changed” is when what the source says last differed from the previous check. A re-check that finds nothing new updates the first date only. Sitemaps and structured data use the “last changed” date.
How changes are verified
- Every check is logged with the source URL, time, HTTP result and a fingerprint of the relevant text.
- If a fetch fails, or the page loads but a section we found before has vanished, we keep the last verified information, retry with backoff, and flag the park after repeated failures. A failed extraction never replaces good data with blanks.
- Large or unusual changes (a fee doubling, hours disappearing, a park dropping off the official list) are held and re-checked before publishing. Parks missing from an official list are flagged, not deleted.
- Renamed parks get a new address with a permanent redirect from the old one. Closed parks stay listed, marked closed, with nearby alternatives.
How we interpret what parks say
- Hours are kept separate for the park, gate, office, visitor center, campground and trails, because these often differ. “Open now” is computed from the official hours in the park’s local time, including sunrise/sunset rules, and only when the hours are unambiguous. Alerts and closures override it.
- Amenities have three states: listed by the park (✓), stated as unavailable (✕), or not mentioned — which we treat as unknown, never as “no.”
- Accessibility is shown as the specific features and wording the park provides, not a yes/no label.
- Distances are straight-line, not driving distance.
- Summaries are written from the park’s own facts (location, size, listed activities and facilities). We don’t copy official descriptions and don’t add claims that aren’t in the data.
What counts as a state park
We include every publicly accessible destination that a state’s park system lists as one of its units: parks, recreation areas, historic sites, natural areas and preserves, beaches, scenic sites and state trails. Campgrounds, beaches and visitor centers inside a park are sections of that park’s page, not separate listings. Areas run by other agencies (wildlife management areas, most state forests, sites run by separate historical commissions) are excluded, and each state page notes what was left out. Parks that cross state lines have one page, listed in each state.
Some pages have too little verified information to be useful yet. They stay in the directory and state lists but are hidden from search engines until more is known.
Photo policy
We only use photos that are verifiably in the public domain: works of the US federal government, works whose creator released them into the public domain (including CC0), and works whose copyright has expired. For each photo we keep the source page, creator where known, and the specific evidence of its status — normally the license template on its Wikimedia Commons file page. Photos that need attribution or carry other conditions (such as CC BY or CC BY-SA) are not used, and state-government photos aren’t assumed to be public domain. Photos must depict the park itself. When no suitable photo exists, the page shows a plain illustration instead — never a generated image presented as a photograph.
How we access sources
Our crawler identifies itself as USStateParksBot with a contact address, follows robots.txt, and limits itself to about one request per second per site. If a site blocks automated access we don’t try to get around it; we use the state’s open data instead, and note the limitation.
Maps
Map tiles are self-hosted from OpenStreetMap data processed by Protomaps. © OpenStreetMap contributors. Park boundaries are from USGS PAD-US.