Every entry on this site records a handful of specific conditions about one place, with a count of how many separate accounts support each and the date it was last checked. This page explains where those come from and, more importantly, where they do not.
Finding the places first
A city’s playgrounds are not held in one list anywhere. Municipal parks departments publish their own sites, mapping services hold others, and a large share of play areas exist inside parks that are not indexed as playgrounds at all.
We assemble the list by searching a city area by area rather than as a whole, because a single search returns only the most prominent results and misses the neighborhood sites that people actually walk to. Each result is then checked against public park boundary data so that a play area sitting inside a larger park carries the name people use for it, rather than appearing as an unnamed point.
Places that turn out not to have children’s play equipment — dog parks, skate parks, sports complexes, nature preserves that happened to match a search — are removed at this stage. This is a larger share than you would guess.
Reading what visitors said
For each place, we read visitor accounts and pull out sentences that describe a specific condition: shade over the equipment, a fence and whether its gate latches, a restroom and whether it was open, the ground surface, what the equipment suits.
General impressions are set aside. Lovely park, we go every week is a true statement about someone’s experience and it tells you nothing you can act on. No shade, no bathroom, no parking is three usable facts in one sentence.
This is the labor-intensive part and there is no shortcut to it. It is also the reason this index exists, because the information is genuinely there and genuinely unusable in its original form.
Deciding what to publish
Four rules govern what makes it into an entry.
We grade rather than flatten. A barrier along the street side of a play area is not the same as a fully enclosed one, so it is recorded as partial with the detail spelled out. Reducing both to fenced: yes would be easier to filter and would mislead the person doing the filtering.
We keep disagreements. When accounts split — shaded seating but exposed equipment, a splash pad some visitors found running and others found closed — we record the split rather than picking the tidier reading. Where accounts disagree, the entry says so.
We weight recent accounts more heavily for things that change: upkeep, equipment condition, whether a restroom was open. A complaint about broken equipment from several years ago is weak evidence today.
We publish our own description, never a visitor’s words. Every sentence on this site is written by us from what accounts collectively report. If you want the originals, the mapping services that host them are one click away.
What stays blank, and why
If nobody has described whether a playground is fenced, the entry says we do not know. It does not say no, and the place does not appear when you filter for enclosed play areas.
This is deliberate and it has a cost: our filtered lists are shorter than they could be, and some places that do have fences are missing from the fenced list. We accept that cost because the alternative error is worse. A parent who filters for enclosure, drives out, and finds an open play area beside a road has been failed in a way that a slightly short list does not fail anyone.
The blank fields are also the work queue. Each entry lists what we could not confirm, and those are the things we most want to hear about from people who were recently on site.
What we will not do
We will not infer from what is typical. Most playgrounds have swings; that is not evidence that a particular one does.
We will not rank places against each other. The best park for a two-year-old in July is not the best park for a nine-year-old in November, and a single ordering would hide the distinction that makes the index useful.
We will not characterize people. Some accounts describe who else was at a park. We record conditions of the place — lighting, visibility, broken glass, upkeep — and leave the rest out. Where the only safety-related evidence is an impression about other visitors, we publish nothing.
We will not fill gaps with generated text. A blank field is more useful than a plausible sentence.
How entries change
Conditions change and accounts accumulate. Fences go up, trees come down, splash pads change their season, equipment gets replaced after a renovation.
Each entry carries the date it was last checked, and entries are re-read on a schedule. Corrections from people who were recently on site jump the queue, because someone who tried the restroom door last Saturday knows something no amount of reading will produce.
If an entry is wrong, tell us what you found. That is not a formality — it is how the parts of this index that cannot be read into existence get built.