A multi-model snow forecast site for any US zip code
Built in two days with Claude, the site compares snowfall models, shows how far they agree, and uses the Kuchera ratio and temperature inversions to spot freezing rain.
Credits
- Reddit poster tour_in_a_box Source ↗
Earliest dated source:
The full story
Snowyay was built over two days with Claude ahead of a storm. For any US zip code it pulls forecasts from several snowfall models, shows how much they agree, gives a timeline of when the snow will fall and an average of how much to expect.
It uses the Kuchera method to estimate snow-to-liquid ratios from atmospheric temperatures, and looks for temperature inversions aloft to detect freezing rain. The data comes from Open-Meteo's free API. The poster says Claude handled almost everything, from the API work to the snow calculations and the design.
How it was done
Tools
Themes
Loading themes…
Sources
From a short post by the maker, so the details are as they tell them.
- Reddit: I built a multi-model snow forecast app with Claude just in time for the stormReddit poster: tour_in_a_boxPosted
The maker's post on r/ClaudeAI.
If it disappears: archived copy of Reddit: I built a multi-model snow forecast app with Claude just in time for the storm (opens in a new tab)
- The project: snowyay.com
Linked in the post.