Misty Finnish lake seen from a jetty

Case study · Data & programmatic SEO

Pulahda.fi — every swim spot in Finland, in one service

We built a directory that turns open data into thousands of search-optimised pages for 3,000+ swimming spots — and every one of them works as well on a phone as on a desktop.

The starting point

Finland has thousands of swimming beaches and spots, but the information is scattered: the LIPAS registry, municipal sites, environmental water-quality reports, weather data. There was no single place to find your nearest beach, its facilities and its water quality. Pulahda.fi fills that gap — and does it so Google finds every location.

What we built

One browsable service for three categories: swimming beaches, winter and ice swimming, and public saunas. Users search by name or browse by city on a map. Every location has its own page: address, type, facilities, season, photos and — where data exists — the EU bathing-water quality class and water temperature from the nearest measuring station.

Pulahda.fi Helsinki hub on a MacBook and the home page on an iPhone

Programmatic SEO: thousands of pages, each earning its place

The site isn’t a handful of hand-made pages. It’s thousands of pages generated programmatically from open data — each with its own clean URL, title and meta description.

  • Over 2,200 location pages (/paikka/…)
  • 306 per-city hub pages (/kaupunki/…) — search terms like “uimarannat Helsinki”
  • 427 locality pages, plus vertical-specific hubs
  • 3,200+ URLs in the sitemap

Titles and descriptions are generated from templates so every page is relevant to its own query — the bathing-water quality class, for example, is pulled straight into the title.

Open data as a pipeline — sources on show

All information comes from open, reliable sources, and they’re credited in the open:

  • LIPAS (University of Jyväskylä, CC BY 4.0) — locations and facilities
  • EEA / WISE — EU bathing-water quality
  • SYKE Hydrology API — water temperature
  • OpenStreetMap (ODbL) — maps
  • Wikimedia Commons — photos, with per-image author and licence

The data refreshes automatically, so water quality stays current with no manual work. Attribution and honest disclaimers — “data may contain errors, always check current information” — are on show, not hidden away.

Pulahda.fi data-sources page listing LIPAS, EEA, SYKE, OpenStreetMap and Wikimedia Commons

AI for fact-checking and enrichment — openly

We use AI as part of the build, and we say so plainly.

  • Content enrichment. Bare registry data becomes readable, human descriptions. An AI-written summary is always labelled as such, with its source shown — we don’t pass machine text off as hand-written.
  • Fact-checking and cross-referencing. AI helps compare sources against each other and surface conflicts, so errors are caught before publishing.
  • Images and visuals. Some background material is produced with AI, deliberately and in a documentary style, never misleadingly.
  • Speed. Building thousands of pages by hand would be impossible; the machine handles the repetition, a human makes the decisions and owns the quality.
Pulahda.fi location page on a phone: an AI-written summary label and an Excellent bathing-water quality card

The principle is simple: the machine does the heavy lifting, a human is responsible for the facts and for what’s true. Machine-generated content is labelled as machine-generated.

Works on a phone — not just “responsive”

Most “find the nearest beach” searches happen on a phone — at the shore or on the way there. So mobile wasn’t an afterthought:

  • The map and its markers load and work full-width on mobile too
  • Filters, quality cards and data tables lay out cleanly on a narrow screen
  • Large touch targets, high contrast, no horizontal scrolling
  • Pages are light (kilobytes, not megabytes) and load in a fraction of a second even on mobile data

Technical quality that holds up

  • Static site: fast, secure, cheap to run
  • Structured data (JSON-LD: breadcrumbs and FAQ) for rich results
  • Google Consent Mode v2, privacy-first (optional categories off by default)
  • Semantic, accessible markup and clear breadcrumbs
  • Creative Commons licences honoured per image

Want something similar?

Pulahda.fi shows what open data, programmatic SEO and disciplined use of AI can do together. If you’ve got data that should become a findable service, let’s talk.

Pulahda.fi contains data from LIPAS (University of Jyväskylä, CC BY 4.0), EEA/WISE, SYKE (CC BY 4.0), OpenStreetMap (ODbL) and Wikimedia Commons. Screenshots: pulahda.fi.