Rain Perceptron
A perceptron learning to separate rainy days from dry ones in a year of real New York weather, using only humidity and pressure. Watch the decision line swing into place as accuracy climbs.
loading data...
Chapter 10 (Neural Networks). Each day is a point (humidity, pressure) labeled by whether it rained. The classic perceptron rule adjusts three weights until the boundary fits, showing what a single neuron can and cannot separate.
Data: Open-Meteo Archive API, daily 2025 humidity/pressure/precipitation for NYC (public, no key).