Skip to content

Use Case and Data

FoehnCast is not a generic forecast portal. It is a decision tool for one rider profile and a fixed set of Swiss kite spots, built around a simple question: where is the best session worth driving to next?

Scope matters here

The project stays intentionally narrow. It ranks a small set of known spots for one rider baseline instead of trying to be a universal weather product.

Use Case In One View

  • Rider

    The baseline user is one rider based in Schwyz with a fixed quiver and weight profile.

  • Decision

    The output is not just a forecast. It is a ranked session choice across multiple spots.

  • Personalization

    Drive time, rider profile, and shore-specific wind context matter alongside raw weather.

  • Scope

    The system is intentionally limited to six Swiss lake spots so the ranking logic stays concrete and testable.

flowchart LR WX[Weather forecasts] --> QUAL[Per-spot quality prediction] SPOT[Spot metadata and shore orientation] --> QUAL QUAL --> RANK[Ranked session options] RIDER[Rider profile] --> RANK DRIVE[OSRM drive time] --> RANK

Rider Profile

Field Baseline value
Home location Schwyz
Weight 80 kg
Quiver 5 / 7 / 8 / 10 / 12 m2

This baseline keeps the ranking problem consistent across the project. The model does not try to optimize for every possible rider at once.

What The System Ranks

The ranking layer combines more than one signal.

Factor Why it matters
Predicted quality index identifies whether a spot looks rideable and how good the session could be
Session duration rewards spots that stay rideable for longer windows
Drive time discounts distant options that are not worth the trip

That is why the system is better described as a trip-planning tool than as a plain weather dashboard.

Spots in Scope

Spot Canton Difficulty Water Ideal wind window
Silvaplana GR intermediate lake 180°-270°
Urnersee UR intermediate lake 150°-210°
Lac de Neuchatel NE beginner lake 200°-310°
Bodensee TG beginner lake 180°-300°
Walensee SG intermediate lake 220°-320°
Thunersee BE advanced lake 220°-330°

These are not example rows. They are the actual fixed spot set used in the current configuration baseline.

Data Sources

Source Role in the project Current status
Open-Meteo primary forecast and archive weather source implemented
OSRM drive-time personalization for ranking implemented
MeteoSwiss possible observation reference for later validation planned

Open-Meteo supplies the forecast data, OSRM already contributes travel-time ranking inputs, and MeteoSwiss remains a possible future reference source rather than an active dependency in the current stack.

Why The Scope Is Fixed

  • It keeps the labeling and ranking problem aligned with one real rider scenario.
  • It makes route-time personalization meaningful instead of theoretical.
  • It avoids turning the course project into a generic weather portal with weak decision support.
  • It keeps later changes focused on implementation maturity rather than on changing the product question.

See Architecture for the current system structure and Feature Pipeline for how the data moves through the stack.