Measuring the complexity of urban form and design

Abstract

Complex systems have become a popular lens for analyzing cities and complexity theory has many implications for urban performance and resilience. This paper develops a typology of measures and indicators for assessing the physical complexity of the built environment at the scale of urban design. It extends quantitative measures from city planning, network science, ecosystems studies, fractal geometry, statistical physics, and information theory to the analysis of urban form and qualitative human experience. Metrics at multiple scales are scattered throughout diverse bodies of literature and have useful applications in analyzing the adaptive complexity that both evolves and results from local design processes. In turn, they enable urban designers to assess resilience, adaptability, connectedness, and livability with an advanced toolkit. The typology developed here applies to empirical research of various neighborhood types and design standards. It includes temporal, visual, spatial, scaling, and connectivity measures of the urban form. Today, prominent urban design movements openly embrace complexity but must move beyond inspiration and metaphor to formalize what "complexity" is and how we can use it to assess both the world as-is as well as proposals for how it could be instead.

A Macroscopic analysis of transport networks: The influence of network design on urban transportation performance

“…Average node connectivity is a useful network proxy, defined as the average over all pairs of vertices of the maximum number of internally disjoint edges connecting a pair of vertices [14]. It is a measure of network resilience: in networks with low average connectivity, circulation is forced through low-permeability choke points, which increases the risk of traffic jams and network disruptions [3]. Another important indicator is the average circuity of a network (or directness) which is the ratio of shortest distance on the network over the Euclidean distance averaged over all origin-destination pairs in the network [5,6,15].…”

“…Tables 5-7 demonstrate the calibration results of the linear function model eqns (2)(3)(4)(5). The coefficients d, e, f and g quantify the reduction in traffic congestions due to an increase/ decrease of the independent variables.…”

“…However, the results of the linear function model (eqn (3)) demonstrate that an increase in AWTCC decreases road congestion at a similar level as does an increase in average road closeness centrality. The results of the linear function model (eqn (4)) demonstrate that the influence of RIA on congestion stays a little lower compared with model eqn (3) and CIA has the highest influence on congestion alleviation compared with an increase in RIA and TIA, while considering cities with over 0.4 km per km 2 cycleway density. Furthermore, an increase in ACCC decreases congestion more effectively than an increase in TIA.…”

“…Urban transportation networks are the distributors of cities for energy, materials and people to specific zones of the city, in the same way as a cardiovascular network distributes energy and materials to cells in an organism [2]. The term complexity for transport networks results in a rich behaviour arising from system connections, interactions with subsets and the dynamic processes where vehicles and people act within a network structure (pattern and configuration) [3]. In recent years, understanding the structure and dynamics of urban transport networks has been improved through analyses of network topology measures using the mathematical tools of graph theory [4].…”

“…It is also an effective evaluation tool since providing feedback of the system is important for correcting, improving or upgrading urban models before executive plans are drafted [10]. Dynamic system variables such as population and traffic volume account for the state of the system as it changes over time [3]. On the other hand, one can evaluate cities through co-evolution, where humans shape their city and are shaped by the city, thus making topological measures an important proxy.…”