A heatwave settles over the Alps, and within hours a fragile moor landscape has a problem it will only measure tomorrow. A "secret spot" reel does the rounds on Instagram and TikTok, a route gets shared on Komoot and AllTrails, and by the weekend there are drones over a game reserve, campfires in a tinder-dry forest and tents pitched in a protected zone. The people posting are already on their way — often before they arrive. The visitor counter at the trailhead, meanwhile, only clicks once they are standing on it.
That gap — between when pressure becomes visible online and when it becomes countable on the ground — is what Ranger Radar was built to close. It is an early-warning cockpit for ranger and protected-area services, piloted for the Lombachalp above Habkern in Canton Bern, and a sibling to the "Ask the Ranger" assistant we run for the same area.
Two lenses on the same landscape
The first lens is geo-fenced social monitoring. For a defined area, the radar scans public posts across Instagram, TikTok, YouTube, Komoot, Outdooractive, AllTrails, Hikr, news and blogs via a search API. Every mention is auto-classified against a protected-area taxonomy — drone, wild camping, campfire, wildlife disturbance, off-trail, dog off leash, parking, crowding — and tagged with sentiment and a risk level. A configurable alert engine (theme × risk × threshold × time window) turns the noise into a short list of things a ranger should actually act on, each with an acknowledge step. Research finds that social-media signals correlate with real visitor numbers and that a meaningful share of shared routes cross no-access zones — which is exactly why this is a leading indicator, not a rear-view mirror.
The second lens is heat and drought. Using live weather against 1991–2020 climate normals from Open-Meteo/ERA5, the radar detects heatwaves, assesses the water balance and renders the now-iconic warming stripes for the area. It matters because the two lenses spike together: when it is hot and dry, visitor pressure and fire and ecological risk rise at the same time.
Public data, private by design
Both lenses sit on a map built from swisstopo basemaps with official BAFU protection-zone overlays — Wildruhezonen, Jagdbanngebiete, Moorlandschaften — so a signal lands inside the zone it threatens, not on an abstract grid.
Because the subject is people in nature, the design is deliberately restrained. Only public posts are read; no personal data is stored; everything runs on CH/EU-hosted infrastructure over official public data. This is a pilot, and the demo dataset deliberately mixes real sources with clearly labelled representative posts — so the cockpit can be shown honestly without turning a demonstration into surveillance.
A counter tells you the crowd arrived. A radar tells you it is coming.