← Atelier

Kite Advisor

Paused AIScrapingAutomatisierung ↗ Live

Started: Feb 15, 2026

Calculate the optimal kite quiver based on historical wind data, body weight and skill level.

WhatsApp Reality Check Feb 20, 2026

Shared the first version in the Lake United WhatsApp group. Immediately got feedback that the results couldn't be right (thanks Simon). Completely overhauled the calculation model, introduced plausibility tests using recommendations scraped from forums as checks. Results are now somewhat reliable — but I still don't blindly trust the wind data.

Wind APIs deliver grid points, not spot-accurate measurements — you get the coastline, not the gap between the dunes where the kite actually flies. The core data problem that no scraping solves. ERA5 data would be a usable proxy, but the download overhead isn't trivial.