Network Analysis
Definition
Network analysis in GIS examines how people, goods, water, energy, or information move through connected systems represented as graphs of nodes and edges. Edges carry costs—time, distance, tolls, risk, carbon—and constraints such as one-way streets or weight limits. Beyond shortest path, modern analysis considers capacity, reliability, temporal schedules, and disruptions. It models accessibility (how far in X minutes), service territories, flow assignment, centrality, and vulnerability. Clean topology is essential; dangling edges or disconnected components can yield misleading results. Network analysis differs from straight-line metrics because it respects actual paths available to travelers or flows, including turn penalties, transfers, and schedules. It scales from indoor maps to continental logistics networks and from road/rail to water, electrical, and communications grids. Results often feed routing apps, equity audits, and resilience planning.
Application
Cities route emergency vehicles and optimize snowplow beats. Retailers plan delivery windows. Transit agencies design feeder routes. Utilities analyze redundancy in feeder lines. Humanitarian logistics compute least-cost supply chains to reach remote communities. Public health evaluates clinic accessibility via realistic walking and transit paths rather than as-the-crow-flies distances. Planners test how new bridges or bus lanes change reach, and resilience teams simulate failures to identify single points of weakness.
FAQ
What’s the difference between shortest path and least-cost path in networks?
Shortest path minimizes distance; least-cost path minimizes a generalized cost (time, tolls, slope, reliability penalties). In practice, time or reliability is usually more relevant than pure distance.
How do schedules change network results?
Time-dependent edges (e.g., transit GTFS) shift reachable areas by departure time. Accessibility at 8 a.m. can differ dramatically from midnight; the model must incorporate headways and transfer penalties.
Why do we need topology validation before analysis?
Disconnected nodes, incorrect one-way flags, or snapped geometries break routing. Automated QA (connectivity, turn tables) prevents silent errors that produce unrealistic routes.
Can equity be evaluated with network analysis?
Yes—compare travel-time catchments to jobs, parks, or hospitals across neighborhoods and demographics to reveal service disparities and target investments.