Coding and AI classes in Motherwell

The learner downloads OpenStreetMap data and keeps the public streets inside a rectangle over Motherwell: 96.9 km, of which 14.8 km carry one-way rules, including roundabouts. Each street becomes one arrow for a one-way street or two arrows for a two-way one. Python then finds the strongly connected components: groups of junctions where every one can reach every other and back. A classic way to do it, Kosaraju's algorithm, runs one search along the arrows and a second search against them.

At first sight 136 points look trapped, but 121 of them sit near the edge of the rectangle. Cutting the map at a straight line chops streets in half, and a one-way street that leaves the box looks like a dead end with no way back. That is a problem created by the analysis, not by Motherwell. Only 15 interior points remain, in three tiny pockets: one 168 m unclassified road that the map lets you leave but not legally enter, and two short pieces of about 15 and 19 m on a tertiary road. Pockets like these are usually a real restriction such as an exit-only road, or a tagging slip worth checking on the ground; the graph cannot tell which.

Separating real problems from artefacts keeps Motherwell teenagers in charge of the AI they use, and coding is where that skill is practised in 2026. The longer argument is in why teenagers should still learn to code in 2026.