Concept · 3 steps · NESA SC4-DA1-01

Application of Models

A model earns its keep by making predictions you can test — from calling a result before it happens, to running the cycle that makes a model better, to knowing exactly when a model is about to lead you off a cliff. Three short steps, each with its own toy to play with.

1 Where it comes fromA model's real job is to predict. Set the inputs, let it guess, then reveal the real data and see how close it got. 2 How to do itThe modelling cycle: build → predict → test → refine → predict again. Drag a trend line until it fits. 3 Where it gets trickyA model that fits the past can still mispredict — especially when you extrapolate far beyond your data.

← This concept rests on

Scientific Models

→ It unlocks

Closes the Year 8 Data-science Strand