Commuter Flows Research Articles

Physical geography and traffic delays: Evidence from a major coastal city

Traffic congestion is a major environmental and social problem whose causes include urban sprawl, imbalanced home-job distributions, increased car ownership, and lack of public transportation. We focus on a relatively understudied factor: the existence of geographic barriers. We study traffic times and flows in the Boston metropolitan area, a major coastal city with substantial shape non-convexities. We show that natural barriers not only cause additional delays to the trips affected directly, but also worsen downtown congestion for everyone. Additionally, commuter flows between places separated by barriers decrease, generating additional traffic elsewhere. We also find that places next to geographic obstacles suffer from higher risks of congestion, due to their lower traffic-diffusion ability. Policymakers may consider specific solutions for congestion arising from constraining physical geographies.

Browsing for food: Will COVID‐induced online grocery delivery persist?

The COVID‐19 pandemic altered daily activities. Many consumers reverted to online grocery shopping and home delivery. We analyze factors associated with the decision to grocery shop online and whether this will persist post‐COVID using data collected via a representative online Qualtrics panel in the State of New Jersey (N = 1,419). Around half of respondents either decreased in‐person shopping, increased online shopping, or pursued a combination of both. We used factor analysis to decompose attitudes towards the pandemic, finding that attitudinal responses broke down into ‘fearful’, ‘believers’, and ‘deniers’. Binomial regressions were used to analyze patterns of frequency of grocery shopping during the pandemic and changes in behavior during the pandemic. Results suggest that age, gender, ethnicity, educational attainment, having children at home, and attitudes towards COVID‐19 are likely to influence frequency of online and in‐person grocery shopping. Specifically, being 50 years or older is negatively associated with online grocery shopping. Those who deny COVID‐19 were less likely to decrease in‐person grocery shopping. People who had children at home, who had advanced degrees, or who were of Hispanic origin were more likely to increase online shopping and decrease in‐person shopping during the pandemic. While our results suggest that in‐person grocery shopping will return to prepandemic levels, we found that respondents report some increased persistence in online grocery shopping post‐COVID.

Detecting shifts in metropolitan structure: A spatial network perspective

The spatial distribution of metropolitan areas creates a complex system of competing and cooperating economic units. Understanding the size, scale and functional structure of such systems is an important topic in regional science. This paper uses commuter flow data over a 40 year period to analyse the changing structure of the Greater Manchester metropolitan area. We apply a combination of complex network analysis, residual network analysis and spatial network visualisation to detect Greater Manchester’s polycentric structure and identify intra-regional communities. This method is able to identify economic geographies in a highly complex and interdependent commuter network. We comment on the role of administrative boundaries in shaping metropolitan regions and discuss the potential of our work to inform debates on regional governance geographies and local government planning practices.

A network optimisation approach to identify trade-offs between socio-economic and ecological objectives for regional integrated planning

Integrated regional planning aims to achieve sustainable development by considering multiple socio-economic, environmental, and ecological goals. However, determining the synergies and trade-offs between sustainability goals is a challenging task that requires an in depth analysis of the system-level effects of the implementation of planning policies. For example, polycentricity (i.e. multiple urban centres) is a common regional urban planning goal, which has shown to be beneficial for a region’s socio-economic conditions, but may also change the configuration of the settlement network (i.e. settlements connected by roads and traffic). Given the dependency of animal species on well-connected habitat networks, another possible planning goal is to maximise habitat availability (i.e. the total amount of habitat that is accessible for an individual animal). However, changes to a region’s settlement network can have a variety of impacts on the region’s habitat networks. Due to these interactions, it is unclear whether maximising polycentricity and habitat availability are compatible planning goals. To address this question, we developed a mathematical model of interacting settlement and habitat networks in a region of Switzerland. The settlement network model allowed us to predict commuter and traffic flows in our study region under a certain distribution of jobs and people across the municipalities. The level of polycentricity in our region was measured by calculating the hierarchy in the commuter flow network. The traffic flow network was linked to the habitat network, which was used to calculate the mean habitat availability. With multi-objective optimisations, both polycentricity and habitat availability were maximised by changing the distributions of jobs and people. Although both goals could be improved compared to the current situation, there was a trade-off between polycentricity and habitat availability along the Pareto front. Developing the region towards either of the planning goals could be achieved by changing the distribution of jobs and people mainly in mid-sized municipalities and by a strong collaboration between municipalities. Our results increase the understanding of the complex interactions in urban regions and can lead to recommendations for integrated urban and conservation planning.

Life in the slow lane: Unintended consequences of public transit in Jakarta

We study how TransJakarta, one of the worlds largest BRT systems, impacted commuting outcomes in Jakarta, Indonesia from 2002 to 2010. Using planned lines for identification, we find that BRT station proximity neither reduced vehicle ownership nor travel times, and it did not increase commuter flows. Instead, the BRT exacerbated congestion along service corridors. To evaluate welfare effects, we calibrate a quantitative spatial general equilibrium model with multiple congestible transport networks. Counterfactual simulations suggest that implementation improvements, including increasing the quality of expansion corridors, would significantly improve welfare with only modest costs.

Spatially-oriented data, methods, and models to plan transit for reverse commuters

The reverse commuting population, those who travel from residences in the city to workplaces in the suburbs, are largely underserved by existing transit systems. One contributing factor to this limited service are issues in existing transit planning methods; serving the reverse commute with transit requires new methods and data sources that complement decision making based on ridership and revealed demand. We utilize administrative origin–destination data from the Philadelphia area that measures commuter flows across all travel modes, rather than just those who already use transit, to capture the potential demand for reverse commuting transit service. We develop an index to highlight target areas for new services, describe geographies for intervention, and utilize negative binomial regression models to analyze key covariates of the reverse commute. Our approach advances a new methodological framework that employs spatially-oriented analyses and open data and tools to generate actionable insights to better serve reverse commuters.

Accuracy of Regional Centrality Using Social Network Analysis: Evidence from Commuter Flow in South Korea

With the recent exponential growth in inter-regional movements of population and information, there is an urgent need for accurately measuring the connectivity and centrality of cities. This study aims to investigate the differences in centrality between different scales of a dataset and to propose a calibration method to minimize the gap between the measures from the two scales. Although urban and regional centrality is examined by analyzing regional commuting datasets, this study proposes that it should be measured using nationwide data to validate the centrality results. To demonstrate this, the differences in regional centrality between different spatial scales of commuting trips for two data groups are shown: Seoul regional data and nationwide data. In this structure, the centrality levels of the 25 districts of Seoul were calculated for both groups. The results clearly show the differences in the centrality levels of districts in both groups: Seongbuk district ranked 10th in the local dataset but fell to 18th in the nationwide dataset; Geumcheon district ranked 22nd in the former but rose to 9th in the latter. The ratio of inner commuting in Seoul is thus relatively low, and each district has dynamic connections with other provinces. Furthermore, the results of a linear regression analysis, which was conducted on a local dataset to obtain similar results as those obtained using a national dataset, demonstrate the significance of a wide-ranging commuting dataset for regional centrality analysis of a specific region.

Research on the Impact of the Built Environment on the Characteristics of Metropolis Rail Transit School Commuting-Take Wuhan as an Example.

The long-distance commute to school caused by urban sprawl and the car-oriented urban construction model are key factors leading to primary/middle school students being picked up by their parents in cars. Encouraging those students to take rail transit can reduce their dependence on cars. This paper uses a stepwise regression based on rail-transit swipe data to explore the influence of the built environment on rail-transit commuting characteristics in Wuhan, and uses a geographically weighted regression (GWR) model to analyze the spatial heterogeneity of significant influencing variables. The study found that: (1) 60% of students are one-way commuters; (2) 88.6% of students travel less than 10 km; (3) the floor area ratio, bus station density and whether the station is a transfer station have an obvious positive effect on the flow of commuters; (4) whether the station is a departure station has a positive effect on the commuting distance, but the mixed degree of land use and road density have a negative effect on the commuting distance. This research can assist cities in formulating built environment optimization measures and related policies to improve school-age children’s use of rail transit. This is important in the development of child-friendly cities.

A spatially aware method for mapping movement‐based and place‐based regions from spatial flow networks

AbstractCommunity detection (CD) is a frequent method for analysing flow networks in geography. It allows us to partition the network into a set of densely interconnected regions, called communities. We introduce a new technique for including geographical weighting in existing methods for detecting spatially coherent communities. We take a link‐based CD algorithm and adjust it to incorporate geographical weighting. We call this approach geographically weighted community detection (GWCD). Our method is demonstrated on two case studies of commonly encountered flow networks: commuter flows and taxi pick‐up/drop‐off flows. Further, we test different measures of distance for geographic weighting and compare our results with the unmodified CD algorithm. Our results show that GWCD can capture the geographical nature of flow regions, generating spatially smaller and more compact areas than if geography is omitted, and that it can be used to distinguish between different types of movement‐type communities.

Viral geographies: Megaregions as extra‐urban disease catchments

Epidemiologists, medical geographers, and public health experts have long focused on city‐centred data in the description of epidemic occurrences. While incidence is collected locally, it is typically aggregated to first a city and from there to a regional resolution. Urban researchers have long noted that interurban travel flow directly affects urban populations and thus models based on a stable urban entity. Some have proposed that a “megaregion,” based on one or another urban exchange network, is a more appropriate vehicle for the study of dynamic disease events. In this preliminary study, megaregions based on American commuter flow are used to study the early spread of H1N1 Type‐A Influenza in 2009—from three principal cities to surrounding towns and cities. The question was, first, whether these catchments would adequately capture the incidence of viral expansion in a region. Second, the question was whether a “wave‐like” pattern of orderly disease progression would be observed.

The rush of the rush hour: mobility justice for seniors on public transport in Sydney, Australia

ABSTRACT Access to transport for older people is inherently related to the right to lead an independent life and to participate in social and cultural life with dignity. Yet often transport planning strategies prioritise the smooth flow of urban commuters, while people over the age of 60 are at risk of experiencing marginalisation and inequality. In this paper, we consider the mobility patterns of a group of seniors who regularly travel by public transport to Sydney, Australia for more than an hour a day. We argue that geographical perspectives are important for understanding how disincentivising seniors from travelling at rush hour may result in a range of negative health and social outcomes. In this research we consider how seniors in fact may benefit from travelling in rush hour materially, socially, and emotionally. Drawing on the optic of mobility justice we examine the felt intensities of using public transport and the inequalities that manifest in corporeal density. This serves to highlight how agency enables people to go against policy aims to reduce senior travel at peak times and how mobility justice is a useful heuristic to dispute power dynamics in specific transport modes.

Multi-disruption resilience assessment of rail transit systems with optimized commuter flows

This paper studies the resilient performance of urban rail transit systems by extending a linear programming optimization model, which can provide optimized commuter flows with contingency routing under multiple disruptions. The rail transit systems in Singapore and Chongqing (China) are selected as case studies, and the system resilience and the effectiveness of providing bus-bridging services are comparatively studied. Intuitively, failures of the interchange stations would inevitably lead to a significant loss of resilience due to deteriorated network connectivity. However, the Singapore case shows that attacking those interchange stations does not always result in an extensive resilience loss. Comparing with the Chongqing case, it is discernible that the position of interchange stations would affect the system resilience. The numerical simulations reveal that the positive effect of providing bus-bridging strategy, in terms of improving the system resilience, is heterogeneous in different systems, which varies from 14% to 30% on average. As demonstrated in the sensitivity on travel demands, this positive effect is robust, yet dynamic, indicating that there is no one-size-fits-all solution for designing transfer-based recovery strategies in disruption management of rail transit systems. The managerial implications for decision makers are also provided and discussed in terms of infrastructure planning.

Tidal phenomenon of the dockless bike-sharing system and its causes: the case of Beijing

Dockless bike-sharing system, as a flexible and eco-friendly solution to improve urban public transportation, has rapidly expanded in many cities around the world. The higher flexibility of the dockless bike-sharing system produces more significant tidal phenomenon that leads to serious traffic problems. However, as a new travel mode, the spatiotemporal characteristics of tidal phenomenon of the dockless bike-sharing system is unknown. This study proposed a method to quantify tidal traffic patterns of shared bikes in Beijing, the capital and megacity of China, and then applied multinomial logit model to reveal main causes of these patterns. Five traffic patterns were found on weekdays, among which three patterns display extreme convergence and divergence states during morning and evening rush hours. Only three patterns exist on weekends and the tidal traffic phenomenon becomes less intensive but lasts longer. Population is the most decisive factor, which determines the density of total traffic flow. Subsequently, resident-employment ratio controls the direction of commute flows thus causing tidal traffic on weekdays, while land use diversity and factors related to leisure activities are more influential on weekends. With the knowledge of tidal phenomenon of dockless bike-sharing usage, some operational strategies were suggested, such as optimizing the stock of the shared bikes in different time and locations, which will benefit bike-sharing enterprises and the local administrators to mitigate problems caused by tidal traffic and promote the usage and efficiency of dockless bike-sharing system.

Forecasting influenza activity using machine-learned mobility map

Human mobility is a primary driver of infectious disease spread. However, existing data is limited in availability, coverage, granularity, and timeliness. Data-driven forecasts of disease dynamics are crucial for decision-making by health officials and private citizens alike. In this work, we focus on a machine-learned anonymized mobility map (hereon referred to as AMM) aggregated over hundreds of millions of smartphones and evaluate its utility in forecasting epidemics. We factor AMM into a metapopulation model to retrospectively forecast influenza in the USA and Australia. We show that the AMM model performs on-par with those based on commuter surveys, which are sparsely available and expensive. We also compare it with gravity and radiation based models of mobility, and find that the radiation model’s performance is quite similar to AMM and commuter flows. Additionally, we demonstrate our model’s ability to predict disease spread even across state boundaries. Our work contributes towards developing timely infectious disease forecasting at a global scale using human mobility datasets expanding their applications in the area of infectious disease epidemiology.

Impact of income on urban commute across major cities in US

Commuting could be costly, while acceptable costs vary across populations. This way, it might be easier for high-income commuters to accept a longer-distance commute having more flexibility in transportation choices and costs. Previous researches provided a statistical description of the gap in the commuting distance in different groups and modelled transportation choices based socioeco-nomic characteristics, but lacks experiments in measuring the capability of distance in discouraging commuting in different income groups. This paper quantifies the heterogeneity of distance’s sensitivity on urban commuting networks by modelling flow volume among different income groups. We analyze the gravity models of the commute flows in 12 major US cities and describe the inconsistency in high-income and low-income workers’ commute behavior. By comparing coefficients in the best fit models, we find that commute decay with distance in 11 out of 12 cities is significantly slower for high-income groups than for low-income groups. Results could help urban planners and transportation engineers to assess equitability in planning public transit and land use.

Cycling mode choice amongst US commuters: The role of climate and topography

Cycling to work is uncommon in most areas of the USA but relatively common in a particular set of metros and neighbourhoods. Explanations for this spatial heterogeneity often focus on differences in local geography, with some areas being allegedly more suitable for cycling. I estimate the role of topography and climate in determining the share of a metro’s workers who cycle to work and the probability a particular worker chooses to cycle. I combine a USA-wide data set of commute flows with detailed elevation and climate data. I find that climate and topography play essentially no role in explaining cycling mode share across metros. Across workers, the hilliness of a commuter’s route is found to be statistically irrelevant to cycling mode choice.

The Spread of COVID-19 and the BCG Vaccine: Evidence from a Natural Experiment in Germany

Background: Since 1 October 2019, COVID-19 has resulted in nearly 18 million reported cases and 700,000 deaths worldwide. The hypothesis that the Bacillus Calmette-Guérin (BCG) vaccine may offer protection against COVID-19 has motivated clinical trials in Australia, the Netherlands and other countries. The results of these trials will not be known for months.Methods: We use a regression discontinuity differences-in-differences (RD-DD) design in Germany exploiting differential BCG vaccination practices in the East and in the West during 1945-90. Our analysis compares cases per capita in German counties within 100 km of the former border for cohorts born around 1990, when mandatory vaccination ended in the East and the West continued to not recommended BCG vaccination. Findings: We analyzed 154,175 COVID-19 cases by single ages reported in 400 counties in Germany on April 24, 2020. We find that cohorts born shortly before and after the policy change exhibit no significant differences in COVID-19 prevalence to those in the west. Eastern counties report fewer log cases per capita than counties just in the west (estimate -.830, 95% CI: -0.528 to -1.132). However, this difference is larger for cohorts born ten years after the policy change (estimate -1.190, 95% CI: -2.384 to 0.003) than those born ten years prior (estimate -0.595, 95% CI: -0.995 to -0.195). A spatial SIR model with commuter flows between German counties produces a differential of -0.588 (95% CI: -1.048 to -0.127) without reference to the BCG vaccine.Interpretation: We interpret our results as providing evidence against the BCG hypothesis. Instead, our evidence is consistent with discontinuities in commuter flows across the former East Germany border being an important contributor to the discontinuity in COVID-19 cases.Funding Statement: None.Declaration of Interests: The authors declare no competing interests.

Characterizing the Regional Structure in the United States: A County-based Analysis of Labor Market Centrality

Categorizing places based on their network connections to other places in the region reveals not only population concentration but also economic dynamics that are missed in other typologies. The US Office of Management and Budget categorization of counties into metropolitan/micropolitan and central/outlying is widely seen as insufficient for many analytic purposes. In this article, we use a coreness index from network analysis to identify labor market centrality of a county. We use county-to-county commute flows, including internal commuting, to identify regional hierarchies. Indicators broken down by this typology reveal counterintuitive results in many cases. Not all strong core counties have large populations or high levels of urbanization. Employment in these strong core counties grew faster in the postrecession (2008–2015) than in other types of counties. This economic dimension is missed by other typologies, suggesting that our categorization may be useful for regional analysis and policy.

A Robust Mixed-Integer Linear Programming Model for Mitigating Rail Transit Disruptions Under Uncertainty

In this paper, we propose a two-stage robust optimization framework to evaluate widely controllable commuter flow strategies that can help mitigate uncertain disruption impacts on rail transit commuter flows. In the operation-as-usual stage, preemptive control strategies including timetable adjustment, commuter path choice diversion, and commuter trip starting time diversion are planned within budget and other requirement constraints. In the disruption stage, optimal contingency routing is derived to minimize disruption impacts. To this end, a linear programming model that simulates the commuter movement logic under train service and platform disruption events is developed, which is then integrated in a mixed-integer max-min optimization model to evaluate the worst-case impacts on commuter flows arising from a set of uncertain disruption scenarios. We show that the two-stage robust optimization framework can be solved by a cutting plane algorithm efficiently using high performance and parallel computing platforms. Finally, we demonstrate the application of the model in mitigating disruption impacts through computational studies based on an actual rail transit network.

Traffic flow splitting from crowdsourced digital route choice support

Digital technology is fundamentally transforming human mobility. Route choices in particular are greatly affected by the availability of traffic data, increased connectivity of data sources and cheap access to computational resources. Digital routing technologies promise more efficient route choices for the individual and a reduction of congestion for cities. Yet, it is unclear how widespread adoption of such technologies actually alters the collective traffic flow dynamics on complex street networks. Here, we answer this question for the dynamics of urban commuting under digital route choice support. Building on the class of congestion games we study the evolution of commuting behavior as a fraction of the population relies on, but also contributes to, crowdsourced traffic information. The remainder of the population makes their route choices based on personal experience. We show how digital route choice support may cause a separation of commuter flows into technology and non-technology users along different routes. This collective behavior may fuel systemic inefficiencies and lead to an increase of congestion as a consequence. These results highlight new research directions in the field of algorithmic design of route choice decision support protocols to help fight congestion, emissions and other systemic inefficiencies in the course of increasing urbanization and digitization.