Europe Experience Score
The whole formula, on one page.
Six numbers appear on every city and country page. Here is exactly how each one is produced, because a score whose method is secret is a ranking, and rankings for sale are how travel sites lose their readers.
Nature
base 34; nature +26, mountains +22, coast +16, islands +12, winter +10, wild +6; +4 per matching experience (max 3), kinds: walk, wild, water
History
base 34; history +26, sacred +16, architecture +16, art +10, cities +4; +4 per matching experience (max 3), kinds: museum, sacred
Food & table
base 34; food +30, wine +22, festivals +6, cities +6; +4 per matching experience (max 3), kinds: table, cellar
Culture
base 34; art +22, architecture +18, music +16, design +14, festivals +10, sacred +8; +4 per matching experience (max 3), kinds: museum, stage, workshop
Adventure
base 34; mountains +24, nature +18, winter +16, coast +14, islands +10, rail +6; +4 per matching experience (max 3), kinds: walk, water, ride
Family
base 34; nature +14, coast +14, wild +12, history +8, islands +8, winter +6; music & nightlife −10; +5 per experience (max 3) that passes the family rule published at /experiences/family
Authenticity
base 34 + 10; quiet +22; not the capital +8; tagged big-city −8; +6 per experience (max 3) of kind workshop, table, wild, cellar; +2 per recorded place (max 4)
Value
110 − (midpoint of the country's daily cost band − €40) × 0.382
Two dimensions this refuses to compute
The specification this came from lists ten. Eight are computable from the dataset. These two are not, and an approximation would be worse than the gap.
Accessibility
We hold no step-free access, hearing-loop or accessible-toilet data for any place in the Atlas. Deriving a number from something else would be a guess about whether a disabled traveller can get in, and that is the worst thing on this list to guess about. The gap is stated on every place page and on /accessibility.
Romance
Not measurable from anything in this dataset. Any formula would be a proxy — sunsets, coastlines, small hotels — dressed up as evidence, and the number would be doing persuasion rather than description.
Discoverability
A second, separate score, used by Discover Mode and by Beyond the Obvious. It answers one narrow question: how far is this place from being the obvious choice?
Not the capital
A capital is where a first visit goes. Everything else has to be chosen.
Editorially quiet
An editor who knows the region marked it as somewhere that stays quiet in season. This is the one judgement in the score, and it is a person's.
Few curated routes pass through
Our own journeys are a proxy for the obvious circuit. A place none of them reaches is off it.
Not tagged for the famous things
Cities, art and architecture are what a continent is famous for. Nature, wild and sacred are what people find later.
In a country we have written thinly
Where the Atlas itself is thin, the place is very likely thin in every other guide too — which is either a gap or an opportunity, and the content report says which.
What discoverability is not
It is not a crowd measurement. We hold no visitor numbers, no search volume and no occupancy data for anywhere in Europe — every product that sells those is licensed, and inventing a proxy for one and calling it evidence is the thing this project exists not to do.
So the score measures obscurity within this Atlas: how far a place is from the obvious circuit as our own dataset describes it. That is a smaller claim than "undiscovered" and it is one we can actually defend. It is also not a quality score — a high number means fewer people will have told you about a place, not that it is better. 109 of 319 places score 70 or above.
How the map is drawn
Open geographic data, hosted by us, with the licence written down before the data was downloaded. No map account, no key, no third-party tile server.
Where the land comes from
Natural Earth, at 1:110 million and 1:50 million. Public domain: the licence says in as many words that no permission is needed and no credit is required. We credit it anyway, because a reader looking at a border is entitled to know which dataset drew it.
How it gets here
A script fetches it and records the SHA-256 of the exact bytes; a second clips it to Europe, simplifies it to three levels of detail and writes the result into the repository. Both are re-runnable, and the build fails if what is committed is not what the pipeline produces.
What it is not
Natural Earth is a cartographic source, built to look right at a stated scale. It is not a legal or authoritative statement of where a border runs, and at the scale of a continent a border is a line several kilometres wide. Do not read a disputed frontier off this map.
Two countries have no shape
Monaco is 2 km² and Vatican City is 0.44 km². At 1:50 million neither has a polygon at all, so both are drawn as a ringed point. Inventing an outline would have been easy, and would have been a lie about a measurement.
Regions are groupings, not boundaries
We hold which region a destination belongs to. We do not hold region geometry — the dataset that has it carries conditions nobody here has accepted — so a region is drawn as its own destinations with its name at the middle of them, and never as a line.
What the scores are not
- Not measurements. They are computed from our own editorial tags. If we tag a city wrongly, its score is wrong, and the fix is to fix the tag — tell us.
- Not quality. A 94 for Adventure means the place is about adventure, not that it does it better than a 71.
- Not for sale. No business, tourism board or partner can change a score. There is no mechanism to; the numbers are recomputed from the dataset on every build.
- Not capped at 100. The ceiling is 97 and the floor is 12, so nothing is ever perfect and nothing is ever worthless.
Why value is different
Five dimensions are structural — they come from tags. Value is the only one anchored to something outside our own judgement: the midpoint of the country's daily cost band, which is published on every country page and can be argued with directly.