A single average temperature is one of the most misleading numbers in any city profile. Tell someone a place sits at 21 degrees and they picture a gentle, unremarkable warmth, the kind of weather you stop noticing. But an average is a midpoint, and a midpoint can be reached from opposite directions. Two cities can share that same 21 and feel nothing alike across a year.
Consider two coastal towns, both averaging 21 degrees annually. One stays close to that figure in every month, mild in January and only pleasantly warm in August. The other reaches 21 by combining cold, grey winters with fierce, draining summers, the two extremes cancelling out on paper. The average tells you almost nothing about which life you would actually be living. That is the problem a climate score has to solve.
Why averages mislead
The mean hides everything that matters about how weather is felt: how far the temperature swings, how long the uncomfortable stretches last, and what the air is doing while the thermometer reads whatever it reads. A responsible climate measure has to look past the headline number to the distribution underneath it.
The factors that change how a climate feels, well beyond its average:
- Variance, meaning the spread between the coldest and warmest months and the size of the daily swing within them.
- Sunshine hours, which shape mood and daily life as much as temperature does, and which two places at the same latitude can differ on sharply.
- Humidity, which decides whether 30 degrees is a dry, bearable heat or an airless one, and whether a mild winter feels crisp or damp and penetrating.
- The length and timing of the difficult season, because a short sharp summer is a very different prospect from five months of it.
Mild is not a temperature. It is the absence of extremes, and you cannot read the absence of extremes off an average alone.
This is why we are wary of any city comparison that leads with a single climate figure. It flatters places with violent seasons and undersells places that are genuinely, reliably comfortable. The honest version is more work, but it is the only version that survives contact with a real year on the ground.
Turning decades of data into one number
The challenge is that climate, unlike cost or visa rules, has to be reduced to something comparable. Our intake asks each person what they actually want, whether that is dependable warmth, four distinct seasons, lots of sun, or simply an absence of the cold or damp they are leaving behind. To match a city against that preference, we need a single normalised value, a climate factor between zero and one, where one is a near-perfect fit and zero is a poor one.
We build that number from decades of historical weather records rather than a single recent year, which smooths out the freak summers and unusually warm winters that would otherwise distort a short sample. From that long record we pull the components that actually shape lived experience: the spread of monthly temperatures, the daily range, sunshine hours, humidity, and the length of the seasons that people tend to find hard. Each is scored against the preference the person gave us, then combined into one figure.
The result is deliberately not a measure of whether a climate is objectively good. There is no such thing. A score of 0.9 means this climate closely matches what this particular person said they wanted, and the same city can score 0.9 for one partner and 0.4 for the other. That divergence is useful. It tells a couple early that they are picturing different skies, and that the climate conversation is one they need to have rather than assume they share.
What the score will never do is collapse back into a single average and pretend that 21 degrees means one thing. The whole point is to keep the texture that the average throws away, so that when a city rises to the top of your shared list, you already have an honest sense of what its Januarys and its Augusts will ask of you.
