Geography and Traffic Quality

Beyond the country chart there are five breakdowns that answer questions a country chart cannot: where exactly, from which network, how good is that traffic, and at what hour of the visitor's own day.

Detailed geography and the traffic quality panel are a paid capability. When the plan does not include them, the panels are replaced by an upgrade prompt and the underlying endpoints answer with a plan restriction rather than with empty data.

The five breakdowns

Regions. Subdivision code, GeoNames identifier, clicks and unique visitors. The identifier is what makes "Bavaria" and "Bayern" one row instead of two.

Cities with coordinates. Each city carries a point and an accuracy radius. Read the radius: a point with a twenty-kilometre radius describes an address range, not a person.

Networks. Autonomous system number, provider name, the proxy type of that range, total clicks and human clicks in separate columns. The gap between those two columns is usually the whole story.

Traffic quality. Visits split by visitor type, by proxy type and by quality score band.

Local hour. Activity by the hour of the visitor's own time zone at the moment of the click. This is computed when the event is written, because the storage engine can only convert to one fixed time zone and every row here has a different one.

Bots are counted here on purpose

Three breakdowns include automated traffic even when "exclude bots" is on: Filtered, Traffic quality and Networks. Excluding bots from them would empty out exactly the panels you opened in order to look at bots.

Everywhere else the "exclude bots" filter behaves as it always has, and automated traffic never consumes the tracked-click allowance of a plan.

Place names follow the reader

An event stores an identifier, never a localised name. The name is resolved when a report is read, from the dictionary of the reader's own language — so one stored row reads as "Berlin" for one colleague and "Берлин" for another.

Names are available in eight of the ten dashboard languages. Italian and Korean fall back to the English name rather than to an empty cell. The dictionaries stay on the server: they are megabytes, and nothing of the sort is sent to a browser.

The language comes from your account setting, not from a request parameter. A parameter would let one person pin someone else's language into a cache shared by the whole workspace.

What is stored and what is not

• The visitor's IP address is not stored. It lives inside one function long enough to produce a daily fingerprint — a salted hash rotated every night — and to look up the address range.
• The postal code is resolved by the database but deliberately not recorded: in dense housing a postal code is close to an address.
• Coordinates are stored as whole ten-thousandths of a degree. Extra decimal places would create an appearance of precision that does not exist.
• Every row records which source answered it, so a chart never mixes precise and approximate data silently.

When the geo pack is unavailable

Redirects keep working and analytics keeps recording. The resolver falls back to the location data the CDN already provides and marks those rows accordingly. City identifiers and accuracy radius are missing from such rows, and the interface says the names are not localised for them.

Data sources and attribution

Location answers come from licensed commercial databases, packed into a format built for edge lookups. The raw database files are never published, redistributed or shipped inside the application: only derived answers about a specific address leave the infrastructure.

City and region names come from the GeoNames geographical database, used under the Creative Commons Attribution 4.0 licence. This attribution is required by that licence and is also published on the privacy page.

Range-type answers — data centre, VPN, public proxy, TOR, web anonymiser — come from a commercial IP intelligence database under the same terms: use inside the service is licensed, redistribution of the file is not.

Controlling detail and retention

Everything above is a setting, in Settings → Visitor data and privacy:

| Level | What is stored | | --- | --- | | Full | Country, region, city, coordinates, accuracy radius, time zone, local hour | | City without coordinates | Country, region, city, time zone, local hour | | Country only | Country and continent |

Separate switches control the visitor's network, the full referring address and traffic-type detection. Retention is a number of days written into each event: the storage engine deletes the row when it expires, which is different from a plan limiting how far back you may look.

Changing any of these takes effect on new clicks within seconds — the settings are pushed to the edge cache of every link and page immediately, rather than at the next cache expiry.