Study finds the hidden forces shaping urban travel

Scientists at Tokyo Tech have developed a groundbreaking metric to measure the complex spatial patterns of the 92 most populated cities in the world. The "inness" metric measures the strength of connection of a city center to its surrounding areas, and helps us understand the effects of rapid urbanization across the globe.

Figure 1. Connectivity metric of individual cities

This figure plots the average and standard deviation of normalized "inness" of the cities in the study. Cities in the LL group, such as Tokyo, have long roads, small geographical restrictions, good connectivity with surroundings, and developed infrastructure. HH cities, such as Calcutta, have short roads, limited connectivity, and limited infrastructure development, forcing traffic through the city center. LH cities are a mix, with low average values and high standard deviations, indicating geographical constraints as in the case of Mumbai.

The empirical study of the spatial distribution of urban land-use and street topology used 472,040 travel routes in 92 cities in all 6 inhabited continents of the globe to understand urban dynamics and associated infrastructure. Urban roads are key building blocks of cities. Past research has shed light on the statistical characteristics of street networks. However, there has been limited understanding of how the city, through its road structure, shapes the movement of people and the flow of socio-economic activity.

This study fills this gap by investigating the competing effects a city can have on the traffic flow—both pulling it in towards its center, and pushing it out to peripheral ring roads. The researchers used randomly sampled origins and destinations to create routes consisting of a series of connected roads, with data from the OpenStreetMap database. By splitting the data into shortest versus fastest routes, the study could the discriminate between physical and functional effects of the road structure.

The novelty of the research is its development of the "inness" metric to capture the direction, orientation, and length of routes, as well as the implicit infrastructural and socio-economic forces that inform route choices. This metric enhances our understanding of the positive and negative externalities arising from urbanization. Although it is a relatively simple measure, "inness" is able to encode the geometric, infrastructural, geographical, and socioeconomic aspects of urban systems, including the distribution of attractive forces that may be implicit in the sampling of streets. Thus, this study was able to investigate the spatial properties of route choices, including land use and street topology; dynamic factors, including congestion, accessibility, and travel demand; and the collective properties of routes, such as the city-wide influence of some roads (e.g., Broadway in New York).

The average "inness" pattern suggests that most cities have a core–periphery structure with high density of streets in the city center and progressively lower density toward the periphery. This pattern is prevalent in cities in developing countries, suggesting that these cities have relatively underdeveloped road infrastructure. At the other end of the spectrum, cities like Berlin display relatively neutral "inness," which suggests that these cities have uniform density of streets. Cities in between this spectrum, like Rio de Janeiro, have both dense and poor connectivity between locations.

Reference

Authors :	Minjin Lee1, Hugo Barbosa2, Hyejin Youn3,4,5,8, Petter Holme6,*, and Gourab Ghoshal7
Title of original paper :	Morphology of travel routes and the organization of cities
Journal :	Nature Communications
DOI :	10.1038/s41467-017-02374-7
Affiliations :	1 Department of Energy Science, Sungkyunkwan University, Korea 2 Department of Physics and Astronomy, University of Rochester, United States 3 Kellogg School of Management at Northwestern University, United States 4 Santa Fe Institute, United States 5 London Mathematical Lab, United Kingdom 6 Institute of Innovative Research, Tokyo Institute of Technology, Japan 7 Goergen Institute for Data Science, University of Rochester, United States 8 Northwestern Institute on Complex Systems, Northwestern University, United States

• Researcher Profile | Tokyo Tech STAR Search - Petter Holme
• Department of Engineering, School of Information and Communications Engineering
• WRHI – Hohan Petter Holme
• Report on the 1st WRHI lecture
• Biointerfaces Unit
• Institute of Innovative Research
• Latest Research News