Network analysis models movement where it actually happens — along roads, rails, pipes and cables — rather than as straight lines across the map. Built on graph theory, it answers the operational questions organisations live by: fastest route, service coverage within a drive time, optimal fleet tours, best facility locations. It is the analysis family with the most direct line to money saved.
The network dataset underneath
Everything rests on a routable network: edges with costs (time, distance), junctions, one-way rules, turn restrictions, and crucially correct connectivity — an overpass wrongly connected to the road below breaks every route through it. Building and maintaining this dataset is most of the real work.
Costs can be static (length, speed limit) or dynamic (traffic profiles, vehicle-specific restrictions), and utility networks extend the same logic to flow, pressure and phase.
The classic problem set
Shortest path is the primitive; service areas generalise it into travel-time polygons around facilities — the honest replacement for circular buffers. The vehicle routing problem (VRP) assigns stops to vehicles under capacity and time-window constraints; location-allocation chooses facility sites to maximise coverage or minimise weighted travel.
Origin–destination matrices, closest-facility dispatch and accessibility indices round out the standard repertoire.
Where it delivers
Logistics compresses fleet kilometres by double digits with VRP; emergency services measure and defend response coverage; public health quantifies who lives beyond reach of care; retailers define true catchments; utilities trace outages through connected networks.
The recurring lesson: results are only as good as the network. Investment in network data quality repays itself through every analysis that follows.