Most explanations begin with the finished tree: boxes, branches, and unfamiliar vocabulary. Aflora begins one step earlier.
You receive ten observations and try to place one boundary that makes their known outcomes less mixed. Only after you can see why that boundary helps do we give the idea its mathematical name.
The complete lesson lets you move the threshold, inspect every candidate split, watch the tree grow from the same calculation, and trace a new observation to its prediction. The visual, the numbers, and the displayed Python implementation all come from the same model state.
This is the pattern future Aflora issues will follow: encounter a difficult idea, make a prediction, manipulate the mechanism, and leave with a mental model you can use somewhere else.