There are many criteria that can be used to classify transportation networks. Its level of abstraction can be considered with tangible network representations closely matching the reality (such as a road map) while conversely an abstract network would only be a symbolization of the nodes and flows (such as an airline network). Since transportation networks have a geographical setting, they can be defined according to their relative location to main elements of a territory, such as a coastal network. Networks also have an orientation and an extent that approximates their geographical coverage or their market area. The number of nodes and edges is relevant to express the complexity and structure of transportation networks with a branch of mathematics, graph theory, developed to infer structural properties from these numbers.
Since networks are the support of movements they can be considered from a modal perspective, their edges being an abstraction of routes (roads, rail links, maritime routes) and their nodes an abstraction of terminals (ports, airports, railyards). Specific modes can further be classified in terms types of road (highway, road, street, etc.) and level of control (controlled access, speed limit, vehicle restrictions, etc.). Flows on a network have a volume and a direction, enabling to rank links by their importance and evaluate the general direction of flows (e.g. centripetal or centrifugal). Each segment and network have a physical capacity related to the volume it can support under normal conditions (traffic above capacity is labeled as congestion). The load (or volume to capacity ratio) is the relation between the existing volume and the capacity. The closer a network is to its full load (a ratio of 1), the more it is congested. Since capacity is often a theoretical estimation, networks can operate above design capacity. The structure of some networks imposes a hierarchy reflecting the importance of each of its nodes and a pattern reflecting their spatial arrangement. Finally, networks have a dynamic where both their nodes and links can change due to new circumstances.