Compressing networks: when roads become points express
An ordinary road map with traffic is a heap of numbers. Sending them all as is takes too long. But you can reduce the volume without losing the gist. Researchers found a clever solution: first redraw the map so that streets become points, and intersections become bridges between these points. Then street traffic is analyzed like sound: complex noise is decomposed into simple rhythms (the principle of spectroscopy). The quietest, barely perceptible beats are discarded — no one will notice their disappearance. The remaining loud rhythms are packed using entropy coding — a method that squeezes out repeats and keeps only the essentials. Scientists also derived a simple smoothness measure: if traffic on neighboring streets changes smoothly, the map compresses much better. In practice, this gives tenfold compression with almost no loss. Interestingly, the trick of turning streets into points was invented back in 1932, but only now has it started working in full force.
The compression ratio is directly tied to information entropy — a concept introduced by Claude Shannon. The more chaotic the data, the harder it is to pack; smoothness is a sign of order.