Forecast station setup
Choose a mode, difficulty, environment, units, and scenario options.
Fictional weather scene
Start a game to generate a scenario.
The canvas will animate clouds, sunlight, rain, wind particles, and lightning from scenario measurements.
Weather measurements
Inspect the fictional station readings.
Make your prediction
Select a weather type, set confidence, then submit.
Forecast feedback
Feature importance
Importance appears after each submitted prediction.
Custom scenario laboratory
Adjust measurements and test your own fictional observation.
Classification rule builder
Create a simple rule, then test it against the current scenario.
Fictional model comparison
Compare simulated classifiers on the same observation.
| Model | Prediction | Confidence | Reason |
|---|---|---|---|
| Run a model comparison. | |||
Interactive Plotly analysis
Explore trends, performance, confidence, distributions, and the confusion matrix.
Observation history
Review fictional records and your submitted forecasts.
| # | Temperature | Humidity | Pressure | Wind | Clouds | Rain chance | Actual | Prediction | Confidence | Result |
|---|---|---|---|---|---|---|---|---|---|---|
| No completed observations yet. | ||||||||||
Achievements
Performance summary
Weather and machine-learning reference
Humidity
High humidity supports cloud and rain formation, but it does not guarantee precipitation.
Air pressure
Rapidly falling pressure can support storm development, especially with strong winds.
Cloud cover
Low cloud cover supports sunny labels. Dense clouds support cloudy or rainy labels.
Features and labels
Measurements are features. Sunny, rainy, cloudy, and stormy are target labels.