Online coding and Python classes in Irvine

The learner downloads OpenStreetMap data over Irvine and keeps every way a pedestrian may use: 350.6 km of streets, paths and tracks, leaving out motorways and trunk roads. The start is the point OpenStreetMap uses to label the town, snapped to the nearest path. Dijkstra's algorithm finds the walking distance to every junction, and each of the 10,769 mapped buildings is given the distance of its nearest junction. At an assumed walking pace of 80 m a minute, 4.8 km/h, the program counts the buildings reachable within each time, and the buildings inside a straight-line circle of the same radius.

The circle badly overstates what is walkable. In ten minutes a walker reaches 292 buildings, only 39.1% of the 746 inside the ten-minute circle. At five minutes the gap is widest: 20 against 152. The share improves with time, because longer walks have more chance to get round whatever blocks the direct line, but even at twenty minutes almost four in ten buildings inside the circle are out of reach. Likely reasons are visible on any map: water, railways and main roads that can only be crossed at certain places, and paths that wind rather than run straight.

The figures rest on assumptions a learner should test: a single walking speed, every mapped path counted as usable, no waiting at crossings, and the choice of start. Change them and the shapes change, which is the point of building the tool rather than trusting one drawing.

An Irvine teenager who can route along real paths will spot a lazy circle in an AI answer straight away, and Python is how they get there. The longer argument is in why teenagers should still learn to code in 2026.