Dani Broitman (p), Itzhak Benenson , Daniel Czamanski

Jean-Claude Thill

There are persuasive arguments for the existence of urban scaling laws. The rank-size rule for cities has been observed in urban systems all over the world at different times. Despite the vast empirical work on this topic, there remain controversial theoretical and empirical issues. There is absence of universal definitions of urban boundaries and the mechanisms that give birth to rank-size rules. Above all, the very concept of urban scaling laws is inherently static. Scaling laws show the configuration status of a system of cities at a specific point of time. However, these are instantaneous realizations of prolonged and at times chaotic dynamics, as was demonstrated by empirical long time series of "rank-clocks" (Batty, 2006). Scaling laws are instantaneous images that reflect a dynamic process at a particular time, but are unable to describe it. In this paper, we argue that the exclusive use of urban scaling laws obscures a rich variety of urban life cycles and the interactions among them. In order to conceptualize the interaction between dynamic life cycles of cities and the static scaling laws associated with them over time, we present an urban model based on firms and simple economic rules that govern their behavior. Competition for workers, salary inequalities and innovation boosted by migrants contribute to create several dynamic scenarios that result in a various rank-size patterns that sometimes reflect only a momentary configuration of an ever-changing system.
Currently, the result are purely theoretical. Their empirical validation is the main objective of the next stage. A wide range of temporal and spatial data on a variety of urban systems is required for such validation. We will discuss possible data sources and their extent and limitations.

Ran Tao , Zhaoya Gong , Jean-Claude Thill (p)

Dani Broitman

Spatial flow data represent meaningful spatial interaction activities such as human migration, daily commuting, and international trade. However, one of the everlasting yet unsolved issue of flow data analytics is the incapability or inefficiency to compute a sizeable flow dataset. This paper focuses on solving computational issues of spatial flow clustering. More specifically, we propose an improvement to the flowAMOEBA approach to enable its computational capability in big flow data analysis. We first discuss the fundamental patterns of flow data distribution, based on which we bring up our solution to accelerate the computation. A case study is conducted with the boosted flowAMOEBA method using an inter-city travel flow dataset in China.

Estelle Mennicken (p), Geoffrey Caruso , Rémi Lemoy

Jean-Claude Thill

See extended abstract

Rachel S. Franklin (p), Eric Seymour

Estelle Mennicken

See extended abstract.

Jean-Claude Thill (p), Yang Zhou

Rachel Franklin

The ease of mobility across the urban environment is known to be a major factor of the spatial organization of the city. It is commonplace to look at accessibility in this space as a structuring element of urban functional areas. In this tradition, the urban space is segmented into nodal regions or communities on the basis of trip origins and destinations recorded to uniformly sized grid cells or Voronoi polygons. However, this approach ignores the role of the layout of the transportation network in forming the regionalization of the urban structure. In this paper, we argue that an effective approach to identify communities in an urban area is by means of its functionally critical elements. The proposed approach starts with the identification of functionally critical nodal points in the city’s transportation system. Then we construct a weighted directed graph based on these functionally critical locations, where each node in the graph denotes a functionally critical location and each edge denotes the presence of travel trajectories between pairs of critical locations; the weight of edges denotes the travel intensity. We introduce recent methods of network science to identify the socioeconomic communities of the urban region. We examine and discuss interesting socioeconomic clusters as well as their variation according to time. Finally we validate the discovered regional patterns by exploring the spatiotemporal patterns of flows on major transportation corridors. We use a big data set that contains all the trajectories of over 11,000 taxis over a month in Wuhan, China. Taxi trajectories have been widely used to study the dynamics of urban mobility (i.e. distance distribution, temporal variability, etc.) and the urban spatial structure (i.e. polycentric distribution, hotspots, borders, etc.). Taxi trajectories are strictly constrained in the transportation network and the accessibility of critical locations in transport network has remarkable influences upon human movement patterns. The results of the analysis suggest that (1) characteristics of socioeconomic clusters are very different from the administrative divisions in Wuhan city. (2) the physical features of the environment, such as rivers, have very limited impacts on socioeconomic communities. (3) The network built from functionally critical locations provides us better ways to understand the regionalization structure of a city, compared to the original node and arc network which is too dense to get the patterns of the whole city.