Network analysis applies graph theory to geographic linear features (roads, rivers, pipelines, utilities) to solve routing, allocation, and connectivity problems. Unlike Euclidean spatial analysis that measures straight-line distances, network analysis respects the constraints of the network: travel occurs along edges (road segments) between nodes (intersections), and costs (distance, time, impedance) accumulate along the path. Core operations include shortest path routing (finding the least-cost route between two points), service area delineation (finding all locations reachable within a time or distance threshold), closest facility analysis, and location-allocation (optimally placing facilities to serve demand).
Spatial analysis typically operates in continuous space -- buffering, overlay, and interpolation assume that distance and movement work in all directions. But much of human activity is constrained to networks: we drive on roads, water flows through pipes, electricity follows wires, data travels through cables. Network analysis brings GIS into this structured, path-constrained world.
A network data model represents the system as a graph: nodes (intersections, junctions) connected by edges (road segments, pipe sections) with attributes encoding cost (travel time, distance, friction). The cost of traversing an edge can be asymmetric (one-way streets, uphill vs downhill flow), time-dependent (rush hour congestion), or multi-dimensional (optimizing for both time and distance). Dijkstra's algorithm and its variants (A*, contraction hierarchies) efficiently find least-cost paths through networks with millions of edges.
Service area analysis extends shortest-path computation to find the frontier of reachable locations within a cost threshold. The 5-minute service area of a fire station includes every road segment and address reachable within 5 minutes of travel along the road network. These service areas are typically irregular shapes reflecting the network structure, not concentric circles. Comparing service areas from multiple facilities identifies coverage gaps and overlaps, guiding facility planning.
Location-allocation problems combine demand modeling with network accessibility to optimally place facilities. Given a set of candidate locations and a demand distribution, the algorithm finds the placement that minimizes total travel cost, maximizes coverage, or satisfies a service standard. This powers decisions about where to locate warehouses, clinics, schools, and emergency facilities -- problems where network distance rather than straight-line distance determines real-world service quality.
No topics depend on this one yet.