← survey home · UK climbs · US climbs · Swiss climbs · Austrian climbs · French climbs

Why this survey exists

On a steep lane above Aberystwyth there is a road sign that says 32%. For years, the same hill carried a sign that said 25%. Same tarmac, same gravity — a quarter steeper, apparently, by decision of a sign-maker.

Everyone who cares about hills has met this problem. A book says one number, a GPS unit says another, the sign says a third, and the rider halfway up has an opinion of their own. I wanted to know which number was true — and it turned out that was the wrong question.

Britain publishes national laser-scanning surveys as open data — aircraft measuring the height of the ground roughly every metre, free for anyone to use. So instead of adjudicating between the sign and the book, I measured the hill — every metre of it, from the public data, with a method anyone can check for themselves.

There is a full circle in that for me. Years ago I authored the Parliamentary Office of Science and Technology's report Electronic Government: Information Technologies and the Citizen (POST, 1998), which argued — among other things — that data the public had already paid to create should be made freely available for anyone to use. This site is a small demonstration of why: open the data, and eventually somebody measures every hill in the country with it.

The answer on that lane — Cefn Llan — was that both signs were arguably right. Over its steepest 2 metres the road is 32%. Over its steepest 100 metres it is 25%. Steepness has no single true value: it depends on the distance you measure it over, and almost nobody who publishes a gradient tells you theirs. That one realisation is the whole site. Every number here states its measuring distance. That is the difference between a measurement and a claim.

I have stood under the 32% sign since, and I still can't tell you how the council arrived at it. What I can say is what would have to be true for it: in our survey this road reaches 32% only over its steepest 2 metres — at 3 m it is already down to 31%, at 5 m to 29% — so if the sign is accurate, it is describing a couple of metres of tarmac, at precisely the scale where measurement is hardest (on a 2 m window, survey noise alone is about ±7 percentage points — ours included). And nobody recorded how the number was made: a survey on the ground, or a short-distance reading taken at face value. That is the whole problem with a bare percentage — and the reason every number here states its distance. You can check every step of the working: the method, the data sources, the exact survey tiles, a machine-readable file per climb. Where a value is less certain — under tree canopy, near a bridge — the page says so instead of pretending.

The survey started with that one Welsh hill, but the method is not specifically British: wherever a country publishes good open elevation data, the same measurement rules apply and the results can be compared. So it rolls on, country by country and region by region: every drivable road is screened, thousands are measured, and only the most notable per region get pages — the rest are measured all the same and will ship in the complete open dataset. Today the site carries 3487 climbs across England, Wales, Scotland and Northern Ireland, and 3905 across the six New England states and New Jersey, New York, Pennsylvania and Hawaii and 3771 across 26 cantons of Switzerland and 4394 across 9 Länder of Austria and 6246 across the Alps, the Vosges, the Pyrenees, the Massif Central, the Jura, Corsica, Provence, the north-west, the north-east and the Centre (70 departments) — the survey's other countries beyond the one it began in. Signs, books and local knowledge aren't the enemy of any of this: they are evidence, recorded and tested like everything else. If you know something a laser can't — a name, a gate, a sign we've missed — tell us.

All of this — the pipeline, the measurements, this site — is the work of one human, enhanced and enabled by Claude Code (Fable), building on open data, open tools and open methods. When these pages say “we”, that is who they mean. Corrections are welcome, and so is scepticism — the point of showing the working is that you don't have to take my word for anything.

Peter Hirst
Aberystwyth, August 2026