Privacy

What is measured

Audience measurement runs on Matomo, hosted in Europe. When a page opens, it records that page's address, the site or search engine you came from, a country and region derived from the IP address, the device type, the browser and its language, the screen size, the time spent on the page, and the outbound links clicked — OpenStreetMap, the town hall website, the directions link.

The IP address is not kept whole: its last two bytes are zeroed before storage — 62.210.0.0, not 62.210.x.y. That stub is what places the visit in a country and a region; it designates nobody. Checked on 25 August 2026 against the stored data itself, not merely against a setting.

What you type in the search box is recorded too, along with how many results it returned — once you stop typing, never letter by letter. The reason is precise: a search that returns nothing points at a track missing from the directory, and there is no other way to learn of it. That box is for finding a stadium: do not write anything about yourself in it.

One extra flag is recorded, and only when it is true: navigator.webdriver, which your browser sets by itself when it is driven by a program rather than by a person. It serves to tell automated agent visits apart from yours. Nothing goes looking for it: the browser announces it, and when it is false nothing is written — a blank does not say « human », it says no signal was seen. No fingerprint is computed, neither of your hardware nor of the way you move the mouse.

None of this names you, and nothing ties two visits together: with no cookie and no stored identifier, you come back tomorrow as a complete stranger.

What it is for: this project asks one precise question — do AI agents read this directory, and do they send people to it? Matomo answers the human half, by sorting arrivals coming from ChatGPT, Perplexity, Claude or Gemini. Without the measurement, the answer would stay a hunch.

No cookie, so no banner

The tracker's first instruction is disableCookies(): no cookie is set, nothing is written to or read from your device in order to measure the audience. That is what makes a consent request unnecessary — the law requires it for accessing your terminal, which this site does not do. The processing itself rests on the controller's legitimate interest (GDPR article 6(1)(f)): knowing whether what he publishes is read.

A browser sending the Do Not Track header is not measured at all: not anonymised, not trimmed — not measured. It costs a few visits; it buys the right to write that sentence without a footnote.

The whole setting is thirty lines long, and it is public: assets/matomo.js. The day a cookie appeared there, this page would say so the same day.

An affiliate link

Venue pages carry one Amazon affiliate link, after the « Access & amenities » section, labelled « Sponsored ». It is a link and nothing more: it displays no image or script from Amazon, sets no cookie, and calls no server until you click it.

If you click and then buy, Amazon pays us a commission. Amazon then knows the visit came from here; what it does with that falls under its rules. Here too we receive only a total: never who bought what.

What this site does not do

• No account, no password, no sign-up.
• No advertising, and therefore no advertising tracker.
• No selling, no exchange of data with a commercial third party.
• No browser fingerprinting, no cross-site tracking.
• No email address collected behind your back: you only give one by writing in, and it is then used to reply to you.

Your location never leaves your device

The « near me » button asks the browser for your position, and the browser asks you. It is used to sort the list by distance, inside the page, on your device. It is sent nowhere: not to the site, not to Matomo, not to the map. Declining costs you the distance sort and nothing else; searching by town or postcode returns the same directory.

What stays on your device

No cookie, but two technical things, neither of which is an identifier:

• an offline cache (a service worker) holding the app's files and the venues, so the directory still works in a stadium with no signal;
• a reparation key in the tab's memory (sessionStorage), which stops the app reloading in a loop after an update, and disappears when you close the tab.

Clearing the site's data in your browser removes both.

Robots, on the other hand, are logged

The domain goes through Cloudflare, where a small program records the passage of crawlers: Googlebot, GPTBot, ClaudeBot, PerplexityBot and the rest. It is the only place they can be seen from — a crawler does not run JavaScript, so no in-page measurement ever reaches it.

That log holds no human visit and no IP address. One line is: the UTC timestamp, the robot's canonical name, its user-agent, the host, the requested path, the page type, the response code, and the country of the data centre the request comes from — which designates a machine, never a person. The code is there to read: logs/worker.js.

What your browser requests from other servers

Drawing a map or an aerial view forces your browser to fetch files elsewhere. Those servers then see, as any requested server does, your IP address and your browser. None of them receives your location, or any data about you from this site.

• GitHub Pages and Cloudflare — hosting and delivery of the pages.
• cdn.matomo.cloud and ronanchardonneau.matomo.cloud — the audience measurement described above.
• tile.openstreetmap.org — the base map tiles, once you open the map tab.
• data.geopf.fr — the IGN aerial views shown on venues that have no photo yet. The images are not stored here; they are requested from the French Géoplateforme as the page is displayed.
• pagead2.googlesyndication.com — the AdSense loader, placed on every page so that Google can review the site. It serves no advertisement: this site's content security policy forbids third-party frames (frame-src 'none'), and an advertisement is displayed inside a frame. That server sees your IP address, as any requested server does; what it does with it falls under Google's rules, not ours.
• Google Maps — only if you click « Directions » or « Opening hours »: you then leave this site, and Google's rules apply.

Contributing means publishing

A photo, a review or a correction is published on the site under the ODbL licence, with the credit you give. You write that credit yourself: a first name, a full name or a pseudonym, as you prefer — it is the only thing about you that will appear.

Two routes, and they are not equivalent. The GitHub form opens a public thread under your GitHub account, visible to everyone. Email lands in a private inbox: your address is used to answer you, never published, never used for anything else.

A contribution can be taken off the site on request. What others have already reused under the licence, however, is beyond the project's reach: that is the price of freely reusable data, and it is better known before you send.

Your rights, and who to write to

The GDPR gives you rights of access, rectification, erasure and objection. For contributions — displayed name, photos, reviews, emails exchanged — this is straightforward: they can be found, corrected, deleted.

For audience measurement, honesty is due: since it keeps no identifier, nothing makes it possible to find your visits among the others, to show them to you, or to erase them selectively. The only way out is preventive: turn Do Not Track on in your browser, and you will not be measured at all.

Data controller: Ronan Chardonneau. To write: ronanchardonneau (at) gmail.com — the address is not a clickable link, for the reason it appears nowhere else in this site's HTML: harvesters would read it. You can also go through the GitHub repository.

If you disagree with how any of this is handled, you may lodge a complaint with the French CNIL.