Pollen over Europe for a month: a crawler that saved the forecast for every hour

The Finnish Meteorological Institute, FMI, runs the SILAM model — a forecast of how pollen spreads across Europe: birch, alder, grass, olive, ragweed, mugwort. But the site only shows the current forecast — maps for the next few hours. And I wanted to see what it looks like over a whole year.

So I wrote a small crawler: once a day it fetched 24 maps — one for every hour — and filed them by date. The first version was in Python; then I started rewriting it in TypeScript.

The script ran for less time than I’d planned: from 12 October to 15 November 2022, about eight hundred and forty mugwort maps. But that was enough to stitch a video.

Mugwort pollen over Europe, 12 October – 15 November 2022, one map per hour. Data and maps: the SILAM model, FMI.

In the video pollen behaves like weather: clouds roll in waves, the wind carries them off, and by November everything drifts south — to Spain and North Africa.

What I took from it: if a service only shows “now”, you can build the history yourself — just patiently save every frame. A full year never came together, but even a month squeezed into half a minute shows what no single map can.