Urban road networks are routinely represented as planar graphs even though overpasses, underpasses, and multi-level interchanges create crossings that no planar embedding can preserve. This study quantifies the distortion introduced by this representational choice across 40 cities. An edge-betweenness-prioritized greedy procedure extracts a maximal planar sub-graph from each network, which is then compared across six metrics. Average shortest-path cost increased in every city, whereas the diameter was unchanged in more than half the sample. The magnitude of distortion scaled with the proportion of edges the planarity constraint excluded, but was unrelated to baseline network scale. Planarization, therefore, requires context-specific assessment.
This paper presents a toll index for Greece, representing the monthly traffic intensity of truck vehicles across urban and interurban highway toll stations. The study period spans the unique recession of the Greek economy during 2008-2016. We employ Autoregressive Distributed Lag models at levels and error-correction form to explore both the short- and long-run responses of manufacturing activity to the toll index. The findings show significant short-run responses in a two-month period, suggesting the utility of the toll index for near-term economic forecasting.
Shared micromobility promises myriad benefits including increasing transit accessibility and reducing car travel but has faced serious challenges including safety and financial viability. We drew upon experts’ opinions to articulate 10 research priorities to support the future of micromobility in the US. Over 70% of 49 experts ranked Infrastructure in their top 5 priorities. Stakeholder groups also expressed different priorities, including greater emphasis on Financing among industry representatives and greater emphasis on Equity among advocates. Findings can inform research funding and dissemination strategies. Further research could examine similarities and differences in priorities across regions and more specific stakeholder groups.
Many transit agencies publish real-time data on vehicle positions, arrival predictions, and service alerts in GTFS-realtime format. While these data provide a wealth of information, to date they have been underutilized in research. In part, this is due to a lack of tooling for reading the bespoke binary format into commonly-used research platforms. This article introduces the `{gtfsrealtime}` R package which allows reading GTFS-realtime into the R statistical programming environment.
Via interviews and oral histories, we examine food access in a low-income, unincorporated community in Dallas County, Texas. We find that car owners and non-owners patronized similar types of stores (primarily big-box stores) for groceries. While owners drove themselves more often than they carpooled as passengers, over 90% of residents used cars to buy food. Yet in-depth interviews and oral histories reveal that measuring transportation resources alone obscures the complexity of community access to food. Despite widespread auto use, many residents had only intermittent access to stores, and some relied on a local food pantry. Some residents also struggled to cook food as they lacked access to electricity and did not own refrigerators or stoves.
Transport planning is, at its core, about enabling people to reach the activities that matter to them. Yet for some people, this goal goes unmet. In this paper, we examine the prevalence and distribution of foregone travel (trips people wanted to make but did not) and transport-related foregone travel (foregone travel attributed to transport barriers) for grocery, healthcare, and leisure destinations in Montreal, Canada. We found that approximately 21% of Greater Montreal residents reported making fewer trips than they wanted for at least one purpose, and 35% of those with foregone travel attributed it to transport barriers. Although most respondents did not report foregone travel, higher rates were reported among lower-income residents and women, while non-white residents showed higher rates of transport-related foregone travel than White residents.
Municipal organic waste programs play an important role in reducing landfill demand and methane emissions, but their performance varies across urban contexts. This study examines the relationship between residential waste collection policy implementation and organic waste collection in the Agglomeration of Montréal between 2012 and 2024. We use data for 33 administrative units and use multilevel models with random intercepts, predicting the total organic waste collection and capture rates, while accounting for socio-demographic characteristics and housing typologies. The results show that organic waste policies targeting single-family and small multi-unit buildings are positively associated with higher total organic waste collection and capture rates, while policies for larger multi-unit buildings show no statistically significant associations. The bi-weekly garbage collection policy is positively associated with higher rates for both outcomes. These findings highlight the importance of organic waste collection policies while suggesting that these policies are strengthened by bi-weekly garbage collection.
The e-scooter is a relatively new transport mode that primarily has replaced other modes. However, e-scooters have also generated some new travel. A meta-analysis was carried out on 192 outcomes of retrospective estimates of the proportion of current e-scooter trips that would not have been carried out of e-scooters were not available. The meta-analysis estimate of induced travel is about four percent, based on the 192 survey outcomes from 95 case-areas. Meta-regression shows that the proportion of induced travel due to e-scooters tends to be higher in areas where travel by car dominates.
Evidence on bus transport accessibility remains limited for rapidly urbanising Asian cities. This study measures bus access to nine essential opportunity types in Hanoi, Vietnam, using open-source datasets. The results show that accessibility was low, highly corridor-based and strongly unequal. At the 60-minute threshold, citywide coverage ranged from only 14.8% of hexagons for sports facilities to 29.1% for retail and grocery, while rural coverage peaked at 23.9%. Population-weighted Gini coefficients remained high, ranging from 0.603 to 0.654 across opportunity types.
Limiting through-traffic and reallocating public space in residential neighborhoods may benefit health and sustainability, yet evidence on public support for such measures is limited. We assessed support for 13 specific measures and neighborhood, mobility, sustainability and sociodemographic correlates in a survey of 1,480 adults from four residential neighborhoods in Berlin. Support was assessed with Likert-scaled items, and associations were examined using linear regression. Support was moderate to high for all measures (61–88%). It was positively associated with active mobility, connectedness to nature, and children in the household, and negatively with car use and age.
Ridehailing activities in Toronto have grown substantially since 2012, prompting concern about the scale of deadheading. Using newly released data from the City of Toronto, we estimate the portion of ridehailing activity attributable to deadheading and examine how it has changed over time, both in vehicle kilometres travelled (VKT) and time spent on the platforms. Our analysis shows that daily ridehailing VKT surpassed 3 million kilometers in 2025, with deadheading accounting for roughly one third of this distance. We also find that the proportion of time drivers spend deadheading has risen since 2022, reaching approximately 48% of total in app time by 2025.
Increases in fuel prices can be driven by a range of factors such as supply chain disruption (e.g. COVID-19 and the 2026 Middle East conflict), general inflation, and changes to government policy settings (e.g climate related emissions reduction policies). Such increases have differential impacts and contribute to social inequity. We calculate a national level index of vulnerability to increases in motor fuel costs at the small area level for Aotearoa New Zealand, drawing on established methods. Those living in “accessible” rural areas near urban centres are more likely to be vulnerable.
Induced demand is usually interpreted as additional capacity reducing generalized cost and producing more travel. When a network exhibits Braess's paradox, an added link raises user-equilibrium cost, and we would thus expect demand to fall. I show this in Braess's original four-node numerical network with linear link costs and inverse demand. In the default case, the added link raises elastic equilibrium cost from 70.14 to 80.00 and reduces demand from 3.66 to 3.33. The reported 143-point sweep maps the Braess region in inverse-demand space; no sign-inconsistent cells appear. The accompanying browser implementation reproduces the reported comparison and sweep.
The ‘critical age of seven’ is widely cited in the media to describe higher pedestrian casualty rates among seven-year-old children in Japan. These children are typically in Grade 1 of elementary school. They often commute to school on foot, generally without parental escort, which exposes them to greater traffic risk. However, pedestrian casualty rates for Grade 1 children have fallen by more than half over the past decade. This study suggests that the concept of a ‘critical age of seven’ may be outdated and could lead to misguided road safety policies.
I estimate the short-run price elasticity of demand for public transit in Belo Horizonte, Brazil using monthly system reports between January 2022 and March 2026, exploiting a 25 percent fare cut on Trunk lines in July 2023 with Circular and Feeder lines as control. The line-level panel with clustered inference and a three-month event study (parallel pre-trends validated, p=0.87) yield a robust estimate near -0.20 across specifications. Heterogeneity across service types is substantial: bus rapid transit responds at -0.42, while diametrical lines carrying over a third of ridership are essentially inelastic at -0.03.
Traffic fatality rates increased more in large US cities than in the rest of the country between 2008-2012 and 2018-2022. Interviews with practitioners from seven high-fatality-rate cities versus seven low-fatality-rate cities suggested that automobile-oriented land use patterns and roadway designs underpinned high fatality rates. While these relationships are well-established, practitioner comments also suggested a need to investigate whether 1) weak institutional support in high-fatality-rate cities prevented changing high-risk system characteristics and 2) more speeding and reckless driving may have contributed to large city fatality rate increases in the 2010s.
Winter weather impacts travel volume, but in cities acclimatized to winter weather the effects are less clear. We study how snowfall impacts travel volume, across trip distances and road types, in the Greater Toronto Area for 96 snow days (2021–2025). Major snowfall days (≥5 cm, n=37) reduce trip volumes by 1.7% and vehicle-kilometers traveled by 2.3%, with effects concentrated on longer trips and higher-speed roads. To understand temporal responses around snowfall, we further examine a stricter subset of temporally independent major snowfall events, defined by grouping consecutive major snowfall days into event units and retaining events with clearly defined pre- and post-event days (n = 23). Mobility is significantly higher before these events than during them (+4.56 percentage points, p = 0.001), while post-event mobility does not show a statistically significant compensatory rebound. These findings suggest that major snowfall produces both suppressive and anticipatory mobility responses: travel declines on major snowfall days overall, while temporally independent events show evidence that forecasted snow may induce pre-event trip-making that offsets or exceeds same-day reductions over a short ±1-day window.
Public Direct Current Fast Charging (DCFC) infrastructure is in its nascency, may not be profitable, and locations may not provide the amenities that consumers want. Using survey data from California this paper explores two topics related to these issues: activities of battery electric vehicle (BEV) drivers and their expenditure while using DCFC. The results show that 44.5% of BEV drivers purchase something while charging. Expenditure in establishments around chargers is $39.82 on average, while expenditure on charging is $10.35 to $16.07 on average. Most spending occurs at cafés, restaurants, or bars; grocery or convenience stores; and department or big-box stores.
This study highlights the vehicle miles traveled (VMT) consequences of zoning, infrastructure investment, and housing planning across California. We developed a statistical model to predict residential VMT per resident at the census block group level in California. Independent variables included built-environment characteristics and housing affordability. The intent of this analysis is to provide a planning-oriented model in which variable selection prioritizes built-environment dimensions subject to direct policy intervention, including density, street connectivity, transit access, and housing affordability, rather than sociodemographic characteristics that, while correlated with travel behavior, lie outside the direct control of land use and transportation planning. Individual models for major MPOs assessed geographical variability in the results and revealed meaningful variation in coefficient magnitudes, suggesting that VMT sensitivity to density, investments in transit, and housing affordability differs across metropolitan contexts. The model provides a practical empirical framework for policymakers evaluating strategies to reduce VMT and related environmental impacts.
Logistic diffusion is often described as passing through early growth, rapid growth, and maturity, but the dates separating these stages, the knees of the curve, are defined in different ways across fields. This note brings several familiar stage-transition rules for the standard logistic curve into one common form. The main point is that each symmetric analytic rule can be written as _t_0_ ± _c/b_, where _b_ is the logistic growth-rate parameter and _t_0_ is the inflection date. Equivalently, each rule implies a fixed percentile window of the asymptote _K_. In particular, the rule based on zeros of the third-derivative implies 21.13% to 78.87% of _K_, and the tangent-at-inflection rule implies 11.92% to 88.08\% of _K_. A small synthetic example also shows how a free three-segment linear regression compares with analytic conventions. This paper clarifies conventions that are often conceived of as conceptually distinct.
