London Underground: Neighbourhood Centrality and Relation to Urban Geography
Weisi Guo
The University of Warwick
Department School of Engineering
Xueke Lu
The University of Warwick
Department School of Engineering
2016 IEEE International Smart Cities Conference (ISC2)
Year: 2016
Pages: 1 - 7
IEEE Conference Publications
Abstract
Transport is important as a means by which people engage in the economy and interact socially. It underpins the business life of the city. Yet, the complexity of urban systems means that we do not always understand how public transport relates to human behaviour and urban features. In this paper, we apply complex network analysis to better understand the London underground and overground rail network and uncover its relation to how citizens make certain lifestyle choices and other urban geography features. We propose to apply Neighbourhood Centrality, which aggregates the network centrality values in a geographic area. We uncover that the neighbourhood eigenvector centrality (influence of the neighbouring rail stations) can explain for up to 79% of the variations in: (1) the age demographics, (2) the choice in transportation mode, and (3) the price and choice of housing. This sheds light on some of the complexities that surround public transport networks and its relation to urban human geography, and can prove useful for the design of future smart city transport systems.
IEEE Keywords
Index Terms—complex network; transport; graph the-ory;
I. INTRODUCTION
Cities and systems within cities are seemingly com-plex and the relationships both within a system andbetween them are not always well understood. In the pastcentury, the number of people living in cities has risenfrom 3% to over 50% of the global population. Alongwith rapid technological development, urbanisation hasled to an increase in the growth of cities and the growthof the complexity in cities. The urbanisation trend haspresented new benefits, but it has also created a setof challenges. Existing studies have shown that boththe benefits and challenges of cities scale super-linearlywith the city’s size [1], and the growing global urbanpopulation certainly exasperates the hidden competitionbetween urban improvements and decay. The question,of how different urban features interconnect and relate to one another has become more pertinent than ever.Transport underpins the business life of the city, andconnects the many separate strands of the life in the city.As personal mobility has increased, the existing urbantransport system has come under pressure from conges-tion and delivering efficient services that meet the qualityof experience expected. One way to better understand thedemand of public transport is by understanding how thefeatures of the existing system relates to (affects and isaffected by) the human geography in cities
Fig. 1. Mapping of the Underground and Overground Rail Network in London: (a) full station and connection map. (b) complex networkversion of the network (not geographically positioned). Node size is proportional to Eigenvector Centrality and darkness of node colour isproportional to Betweenness Centrality. (c) selective network statistics.
IV. CONCLUSIONS AND DISCUSSIONS
Transport is one of the most important aspects ofmodern cities and determines both the mobility of labourand influences the well-being of commuters. Yet, thecomplexity of urban systems means that we do notalways understand how public transport relates to humanbehaviour and urban features. Existing studies havelargely analyzed the efficiency of transport separatelyfrom other urban human geography features .
In this paper, we devise new complex network metricsto suit the understanding of public transport usage.Neighbourhood Centrality aggregates the centrality mea-sure over a small area and the eigenvector centrality isused to measure the influence or importance of a neigh-bourhood. The rationale for this metric is based on thefact that citizens walk to stations of importance in a givenneighbourhood, as opposed to the nearest station. Whenwe correlated the centrality metric against census dataat the ward level, we uncover that the neighbourhoodcentrality metrics can explain up to 79% of the variationsin: (1) the demographic breakdown in age categories, (2)the choice in transportation mode, and (3) the price andchoice in housing. These results shed light on some ofthe complexities that surround transport, and its relationto urban demographics and human geography. In particu-lar, we show a strong bias in demographics towards thoseof working age in neighbourhoods that have importantunderground stations (up to 80%), and a staggering adultto children ratio of 9:1. We also show that the importanceof underground stations in the neighbourhood stronglyaffect the transportation choice, with convex variations inthe percentage that bike to work and strong correlationswith the car ownership and PTAL values. Furthermore,the house price is affected more by the local connectivitypattern of the underground, whereas the percentage offlats and ownership are all affected by neighbourhoodinfluence.
The transferability of this methodology to other citiesis direct and straightforward. Only access to publictransport network data is needed. Yet, what can be doneis less straightforward. For now, all we have observedare interesting correlations (without inferring causality)on the relationship between complex network centralitymetrics and urban human geography factors. Theseresults can prove useful for the design of future smartcity transport systems.
REFERENCES
[1] L. Bettencourt, “The Origins of Scaling in Cities,” Science, vol.340, 2013.
[2] M. Barth ´elemy, “Spatial networks,” Physics Reports, 2011.
[3] S. Porta, P. Crucitti, and V. Latora, “The network analysis ofurban streets: a dual approach,” Physica A: Statistical Mechanicsand its Applications, 2006.
[4] M. Barthelemy and A. Flammini, “Modeling urban street pat-terns,” Physical Review Letters, vol. 100, 2008.
[5] W. Jung, F. Wang, and H. Stanley, “Gravity model in the Koreanhighway,” Europhysics Letters (EPL), vol. 81, 2008.
[6] A. Montis, M. Barthlemy, A. Chessa, and A. Vespignani, “Struc-ture of inter-cities traffic: a weighted network analysis,” Environ.Plan., vol. 34, 2007.
[7] P. Sen, S. Dasgupta, A. Chatterjee, P. Sreeram, G. Mukherjee,and D. Manna, “Small-world properties of the Indian railwaynetwork,” Physical Review E, vol. 67, 2003.
[8] P. Kaluza, A. Kolzsch, M. T. Gastner, and B. Blasius, “Thecomplex network of global cargo ship movements,” Journal ofRoyal Society Interface, vol. 13, Jan. 2010.
[9] R. Guimera and L. Amaral, “The worldwide air transportationnetwork: anomalous centrality, community structure, and cities’global roles,” Proceeding of the National Academy of Sciences(PNAS), vol. 102, 2005.
[10] V. Latora and M. Marchiori, “Is the Boston subway a small-worldnetwork?” Journal of Physics A, vol. 314, 2002.
[11] R. Gallotti and M. Barth ´elemy, “The multilayer temporal networkof public transport in Great Britain,” Nature Sicentific Data, 2015.
[12] M. Rombach, M. Porter, J. Fowler, and P. Mucha, “Core-Periphery Structure in Networks,” SIAM Journal of AppliedMathematics, vol. 74, 2014.
[13] H. So, R. Tse, and S. Ganesan, “Estimating the influence oftransport on house prices: evidence from Hong Kong,” Journalof Property Valuation and Investment, 1998.
[14] K. Thompson and P. Schofield, “An investigation of the rela-tionship between public transport performance and destinationsatisfaction,” Journal of Transport Geography, 2007.
[15] T. Opsahl, F. Agneessens, and J. Skvoretz, “Node centralityin weighted networks: Generalizing degree and shortest paths,”Social Networks, vol. 32, Jul. 2010.
[16] M. Latapy, “Main-memory Triangle Computations for Very Large(Sparse (Power-Law)) Graphs,” Theoretical Computer Science,vol. 407, 2008.
[17] K. Sohn and D. Kim, “Zonal centrality measures and the neigh-borhood effect,” Transportation Research Part A: Policy andPractice, vol. 44, 2010.
[18] V. Bhola, M. K. Grover, M. Sinha, and G. Singh, “Identifyingkey players in a social network: Measuring the extent of anindividual’s Neighbourhood Connectivity,” in IEEE InternationalWorkshop on Business Applications of Social Network Analysis(BASNA), 2010
للقراءة والتحميل اضغط هنا
ليست هناك تعليقات:
إرسال تعليق