Equity of bike infrastructure access in the United States: a risky commute for socially vulnerable populations

This study examines the bike access risk gap (BARG) for commuting in the 50 most populated metropolitan areas in the United States and equips bike advocates with the knowledge and tools necessary to identify the priority areas that need bike infrastructure improvements and the well-connected low-stress bike infrastructure. The analysis (i) examines the average BARGs of metropolitan areas for twelve travel time thresholds, (ii) considers the temporal and spatial disparities of slightly and extremely risky bike infrastructure, and (iii) reveals the disproportionate exposure of socially vulnerable populations to extremely risky bike infrastructure for a journey to work. The results indicate that (i) few metropolitan areas are associated with slightly risky bike infrastructure, (ii) the exposure to extremely risky bike infrastructure becomes more likely as commute travel time increases, and (iii) African Americans, Hispanics, low-income, and carless households are disproportionally exposed to extremely risky bike infrastructure and yet are the least prioritized in urban planning and bike infrastructure investments. The findings offer insights for identifying areas in which constructing low-stress bike infrastructure on or near high-stress bike infrastructure narrows the BARG.


Introduction
On a global scale, the United States is neither ranked in the top 10 as a sustainable country (Berry 2021) nor the top 20 for the most bike-friendly cities (Jahns 2021). The car-dependent nation is 40 years behind Europe due to the sheer lack of connected bike networks and supporting infrastructure for individuals and abilities (Wells 2022). The success in Europe is a case in point of the proven willingness to work toward and maintain sustainable commutes and infrastructure. Sustainability advocates in the United States are optimistic that the enactment of the $1.2 trillion 'Infrastructure Investment and Job Acts (IIJA),' which was signed into law by President Biden 15 November 2021 (U.S. Congress 2021), might set the sustainability infrastructure dialogue in motion. In the bill, the transportation alternatives program, including the recreational trails program, were beneficiaries of a funding increase from $850 million to $1.44 billion per year. This is an opportunity to secure fast-tracked prospects for economic prosperity and social inclusivity for members of the society, particularly as the funding gap has long existed and further created a social divide among citizens and their communities. The funds' appropriations to state and local governments might serve as enablers for equitable, sustainable infrastructure if the distribution of the funds is from a purview of equity. Providing an equitable distribution of funds and infrastructure, however, necessitates the employment of knowledge, tools, and policies to identify the 'priority areas' that are in urgent need of sustainable infrastructure improvements.
Biking as a sustainable commuting alternative can be resolute in addressing climate change resiliency, flexibility, productivity, and technology advancements that are holistically premised on the coupling of social, economic, and environmental dimensions (Wang et al 2018, Milani et al 2021. Despite established benefits at both the individual and the community level (Leister et al 2018), the share of biking for all trips in the United States is 1.1% compared to 26.8% in the Netherlands, 11.5% in Japan, and 9.3% in Germany (Goel et al 2022). Much of the difference is due to the lack of access to connected, stress-free, risk-free, and conventional bike infrastructure in the United States. Studies revealed a positive association between bike ridership and highly connected infrastructure network (Rietveld and Daniel 2004, Krizek et al 2009, Schoner and Levinson 2014, Xu and Chow 2020, dedicated bike paths and lanes (Pucher et al 1999, Parker et al 2013, Schneider 2013, Burbidge and Shea 2018, and the existence of stress-free and risk-free bike infrastructure including separate bike lanes, off-road routes, and bike storage (Cervero et al 2009, Handy andXing 2011). A pertinent example is the research on 43 large United States cities providing evidence that bike infrastructure will be used if built (Dill and Carr 2003). Yet, comparing bike infrastructure on a global scale, U.S. cities rank 26th to 32nd in investment and infrastructure quality scores, fifth in infrastructure scores, and 70th in safety scores. Overall, San Francisco is the top U.S. city with the ranking of 39 on the 'Global Bicycle Cities Index.' This is a policy product, as Freemark et al (2022) expressed the call for engagement between municipality leadership and community members to determine the most viable bike infrastructure policies that connect land use and transport.
As empirical evidence confirmed the effectiveness of bike suitability in increasing ridership (Buehler and Pucher 2012), metrics such as 'bicycle safety index rating' (Davis 1987), 'bicycle stress level' (Sorton and Walsh 1994), 'bicycle suitability rating' (Davis 1995), 'bicycle level of service' (Landis et al 1997, Jensen 2007, and 'bicycle compatibility index' (Harkey et al 1998) have been proposed to evaluate the comfort and safety of bike routes. Of relevance to our study is the level of traffic stress (LTS) classification scheme with four levels of traffic stress (i.e., LTS 1, LTS 2, LTS 3, LTS 4) ranging from lowest stress to highest stress. The LTS metric incorporates bike route characteristics (e.g., number of lanes, speed limit, bike lane blockage, pocket bike lanes) and mimics 'strong and fearless,' 'enthused and confident,' 'adults interested but concerned,' and 'children interested but concerned' classes of traffic stress tolerance (Furth et al 2016). This has been adapted to (i) examine the fractions of commutes connected by the LTS level in the city of San Jose, CA (Furth et al 2016), (ii) measure the 'access gap' to employment by comparing low-stress access with high-stress access in Minneapolis-St. Paul, MN, Miami, FL, Seattle, WA, and Washington, D.C (Murphy and Owen 2019), and (iii) rank the 50 most populated metropolitan areas in the United States for their access to employment by bike at different levels of traffic stress (Owen and Murphy 2020).
The overarching goal of the current research is to equip bike advocates and practitioners with the knowledge and tools necessary to identify the priority areas that need bike infrastructure improvements. Our study, not intended to be a planning or policy piece, offers insight into the social and spatial inequitable distribution of bike infrastructure in the United States. We naturally progress previous research by introducing bike access risk gap (BARG) as an index to measure the gap between bike access to employment by low-and high-stress routes for the 50 most populated metropolitan areas. More definitively, our objectives form a triad overlaying temporal disparity, spatial disparity, and social equity of bike infrastructure. First, the disparity of bike infrastructure is achieved by measuring the BARG and ranking bike infrastructure from best to worst over twelve travel-time thresholds. Knowledge of the disparity of bike infrastructure offers insights into the level of risk in bike infrastructure over increasing commute times and the metropolitan areas ranking of employment accessible by bike. Second, the spatial disparity of bike infrastructure is executed by assessing and comparing the metropolitan land areas covered by slightly and extremely risky infrastructure for each time threshold. This analysis is paramount to understanding the land areas covered by the different risk classifications with increasing commute times and land risk coverage areas for each metropolitan area. In essence, knowledge is gained on the spatial distribution of the BARG between and within metropolitan areas. Third, social equity of bike infrastructure assesses the BARG experienced by the socially vulnerable groups comprised of African Americans, Hispanics, low-income households, carless households, and the elderly. Here, we explore the disproportionate exposure of socially vulnerable groups to extremely risky bike infrastructure. This is a revelation of the inequity in access that stifles economic prosperity, which in turn affects quality of life. The bike access evaluation is conducted using data from the 50 most populated metropolitan areas in the United States. We employ a worker-weighted measure that converts bike access data from the census block level to the census block group level based on data calculated by the University of Minnesota Accessibility Observatory. Data and methods are established next, results and discussion follow, and conclusions are described thereafter.

Data and methods
Data on bike access to employment opportunities are calculated by accessibility observatory at the University of Minnesota for four levels of traffic stress at the census block level (Owen and Murphy 2019). The geography definitions are borrowed from the U.S. Census Bureau's Topologically Integrated Geographic Encoding and Referencing program, the distribution of employment and worker populations are extracted from the U.S. Census Bureau's Longitudinal Employer-Household Dynamics program, and the bike network is obtained from OpenStreetMap (OSM). Each street segment and intersection for the bike network is assigned a LTS adapting a set of hierarchical classification rules. It considers (i) generic paths and crossings that allow bikes, (ii) footpaths and sidewalks that allow bikes, (iii) separated bikeways, (iv) roadways with a bike lane, (v) shared busways and lanes, (vi) OSM tag 'highway' is 'residential' or 'living_street,' and (vii) speed limit. For example, roadways with a bike lane and speed limit of equal or less than 25 mph are assigned LTS 1, and roadways with a bike lane and speed limit of equal or more than 35 mph are assigned LTS 4. Each LTS level is a maximum-allowed traffic stress level. LTS 4 routing includes all street segments and intersections except limited-access facilities, LTS 3 routing includes LTS 1, LTS 2, and LTS 3 street segments and intersections, and LTS 2 routing includes LTS 1 and LTS 2 street segments and intersections. Murphy and Owen (2019) documented roadway and intersection attributes for the bike LTS assignment in detail.
The cumulative opportunities measure (Ermagun 2021) is employed to calculate the number of employment opportunities reachable by bike at different levels of traffic stress for twelve 5 min travel-time intervals ranging from 5 min to 60 min. It is calculated by conducting the Dijkstra shortest-path routing for traveling from the centroid of each census block to the centroid of all census blocks lying within 20 km of the boundary of the origin with the bike speed assumption of 18 kph. The calculations are performed using OpenTripPlanner, and guarantee all employment opportunities are reached within a 60 min travel time. The data prepared at the census block level is converted to the census block group level by employing worker-weighted measures. This helps merge the bike access data with the socioeconomic and demographic variables extracted from the 2019 American community survey (ACS) 5 year estimates for conducting the social equity of bike infrastructure analysis.
Here, we introduce the BARG index, a similar index to the modal access gap (Maharjan et al 2023, Janatabadi et al 2022, to measure the standardized difference between high-stress bike access (A LTS 4 ) and low-stress bike access (A LTS 1 ) to employment opportunities. Mathematically represented in equation (1), BARG is: The BARG index ranges from 0 to 1 and can be measured for different travel-time thresholds. A departure of BARG from 0 indicates a higher likelihood of exposure to risky bike infrastructure when reaching opportunities. Table 1 summarizes corresponding BARG ranges, ratios, and interpretations. The transition from the LTS to safety risk is not improper and is evidenced by research identifying the association between traffic stress and crash severity and validating the LTS as a proxy for safety risk (Chen et al 2017).
Beyond assessing the BARG for each metropolitan area, we examine the land area covered by the extremes of the risk classifications-slightly risky and extremely risky for all travel time thresholds. The land area of each risk classification is divided by the total area for each metropolitan area. We also examine access inequity of vulnerable groups of people that comprise African Americans, Hispanics, low-income households, carless households, and the elderly. This analysis is conducted by comparing the share of the population for a specific vulnerable group (e.g., African American) in extremely risky areas with the share of the population for the same vulnerable group in non-extremely risky areas.
The strength of BARG compared to 'bike suitability' indices is twofold. First, the classification rules of the LTS metric employed here follow a simplification of traditional LTS utilizing roadway classification with measurable characteristics of the infrastructure (e.g., number of lanes, speed limit). This makes assigning the LTS to each street segment and intersection possible at a national level and offers meaning to roadway managers. Second, the BARG index integrates access to valued destinations into 'bike suitability' and offers a 'bikeability' assessment (Lowry et al 2012). This is practical as bike suitability of street segments, and intersections matter when they connect cyclists to valued destinations (e.g., employment, supermarkets, pharmacies). The BARG index reflects on the comfort and safety of an entire bike network accessing valued destinations. The limitations of BARG lie within its strengths. First, the LTS metric is limited to measurable characteristics of the infrastructure and leaves out perceived stress and physiological stress. Second, the current form of the LTS metric, albeit expandable for future research, does not consider the existence or absence of bike amenities (e.g., bike rooms, bike shelters, bike racks). Third, the LTS metric is constructed by OSM data that is susceptible to consistent coverage and tagging quality as OSM is a volunteer-contributed database affected by the knowledge and expertise of the contributors.

Risk classification BARG Ratio Interpretation
Slightly risky 0 ⩽ BARG < 0.25 1.0-1.7 • The number of employment opportunities reached by low-stress bike infrastructure is slightly less compared to the number of employment opportunities reached by high stress infrastructure. • Cyclists are more likely to be exposed to slightly risky bike routes to reach their employment.
Somewhat risky 0.25 ⩽ BARG < 0.50 1.7-3.0 • The number of employment opportunities reached by low-stress bike infrastructure is somewhat less compared to the number of employment opportunities reached by high stress infrastructure. • Cyclists are more likely to be exposed to somewhat risky bike routes to reach their employment.
Moderately risky 0.50 ⩽ BARG < 0.75 3.0-7.0 • The number of employment opportunities reached by low-stress bike infrastructure is moderately less compared to the number of employment opportunities reached by high stress infrastructure. • Cyclists are more likely to be exposed to moderately risky bike routes to reach their employment.
Extremely risky 0.75 ⩽ BARG ⩽ 1 7.0+ • The number of employment opportunities reached by low-stress bike infrastructure is extremely less compared to the number of employment opportunities reached by high stress infrastructure. • Cyclists are more likely to be exposed to extremely risky bike routes to reach their employment.

Results and discussion
Results of the bike infrastructure analysis are broken into three areas of temporal disparities, spatial disparities, and social equity, by measuring the BARG for twelve travel-time thresholds across 50 metropolitan areas. Figure 1 visualizes the average of BARG over census block groups in each metropolitan area for twelve travel-time thresholds. The current state of the bike infrastructure based on the BARG reveals that an exceptionally substantial portion of metropolitan areas are covered by a wide gap between low-stress and high-stress bike access to employment opportunities. An implication is that most bike commuters are more likely to be exposed to extremely risky and high-stress bike routes to reach their employment. In shorter commutes, there are more combinations of risk levels. As the commute time increases, the risk combinations reduce until only extremely risky infrastructure becomes dominant. Slightly risky bike infrastructure is minimal and only available in noticeably short commute times of 5 and 10 min. In excess of 30 min, the bikers have to take extremely risky routes to reach employment in all metropolitan areas. This is due to lack of well-connected, low-stress bike infrastructure. For the top five ranked BARGs, Louisville and Chicago occur most frequently considering the BARG for different commute travel times. In Charlotte, Atlanta, Jacksonville, and Detroit, the BARG is narrow for short commutes, while in Denver, the BARG is narrow for long commutes. This implies a connected network of low-stress bike infrastructure between residential areas and employment in Denver that allows longer commutes and is further backed by the range of infrastructure with 2.7 miles of bike infrastructure per square mile-64.6 miles of paved public paths, 12.33 miles of protected and buffered bike lanes, and 330 miles of other bike lanes (The League of American Bicyclists 2016). In San Francisco and New York, the BARG is narrow for mid to long commutes. In a review of the land use patterns and transport development, there is a prominent level of connectivity of bike infrastructure in residential areas with employment. The same pattern is noticed for Kansas City, which is attributed to the interconnectivity of the varying bike infrastructure types between residential and employment areas (Mid-America Regional Council 2022). The rank of metropolitan areas according to their average BARG for different travel time thresholds is documented in supplementary data I.

Temporal disparity of BARG
Albeit the San Jose metropolitan area offers the largest BARG, as evidenced by figure 1, the city of San Jose has prioritized biking by completing approximately 400 miles of on-street bikeways and 62 miles of off-street trails (City of San Jose 2021). Memphis, another metropolitan area with a large BARG between our metropolitan areas, is shown to have disconnected bike paths and lanes (City of Memphis 2019) between

Spatial disparity of BARG
Analysis of land area coverage of BARG examines the spatial distribution of BARG between and within metropolitan areas. Figure 2 depicts the spatial distribution of BARG between metropolitan areas. This shows the land areas that are covered by slightly and extremely risky infrastructure for twelve commute travel times. The accompanying percentages of metropolitan areas with slight risk and extreme risk area coverages are documented in supplementary data II. The analysis reveals that a narrow BARG is dominant for short commute durations, but exposure to extremely risky bike infrastructure becomes more likely with an increase in commute duration.
As the commute time increases, a lower portion of metropolitan areas is covered by narrow BARG. This is visually represented in figure 2(a). The largest share is associated with San Diego for 10-, 45-, 55-(tie with Denver), and 60 min commutes. Denver has the largest share of land covered by narrow BARG for 15-, 20-, 25-, 30-, 35-, 40-, and 50-min commutes. Providence has the least portion of land covered by narrow BARG for 5-and 10-min commutes at 18% and 7%, respectively. This land area coverage shrinks to 0% coverage for most metropolitan land areas covered by narrow BARG as commute time increases. Seattle has the least land coverage for 15-and 20-min commutes at 0.3% and 0%, respectively. As the commute time increases, the land area covered by large BARG increases. A visual representation is shown in figure 2(b). In 5-and 10-min commute time, Boston has the largest land area with a large BARG. In excess of a 15 min commute, the area of land covered by large BARG is attributed to Seattle, with the land coverage significantly increasing over time. Las Vegas has the least land area with large BARG in 5-and 10-min commute time. For a 15 min commute, Providence has the least land coverage with a large BARG. New Orleans is associated with the least land coverage with large BARG for 20-and 25-min bike commutes. In excess of 30 min, Providence offers the least land coverage with a large BARG. Figure 3 shows the spatial distribution of BARG for a 20 min commute within and between the 50 most populated metropolitan areas. A commute time of 20 min is selected for depiction as (i) it is comparable to bike commute times reported by the ACS and (ii) the average BARG values at the metropolitan areas fall in the extremely risky category beyond a 20 min commute, as shown in figure 1, leaving minimal spatial variation for analysis.
The outcome implies that the bike experience is stressful and not as comfortable as desired for a journey to work. This intuitively defeats cycling for sustainability. Cycling, in comparison to other travel modes, is cost-effective and affordable. Infrastructure costs to support biking are also less than the costs for roadway and public transit infrastructure. By virtue of biking affordability, biking is considered the most equitable travel mode choice and advances sustainability-economically, socially, and environmentally (McQueen et al  2021). Infrastructure investments in the United States, however, have largely focused on highways, bridges, and intelligent transport systems. This remains the case in the IIJA, where biking infrastructure received the least attention compared to other car or transit modes. Biking in the United States is yet to gain steam as a dominant travel mode. Policies on sustainable travel modes to aggressively pursue the major issue of auto dependency and, in turn restrict car use, increase car parking costs, and reduce car parking availability to choke car use for commutes are yet to be adopted. Smaller policies promoting biking promotion, safety, bike parking, and providing bike facilities are central to major active mode plans. The crux of this audacious move to restrict car use is simply to allow more sustainability by reducing traffic congestion, pollution, and carbon emissions.  Figure 4 displays the distribution of the population exposed to different stress levels of bike infrastructure during their 20 min commute. It is evident that the proportion of the population exposed to each stress level varies within and between metropolitan areas. Out of 176 million people exposed to different stress levels of bike infrastructure, 65.7% are exposed to extremely risky bike infrastructure for their 20 min commute. The highest and lowest share of the population exposed to extremely risky bike infrastructure happens in San Jose (87.5%) and Louisville (40.4%). The share of the population exposed to moderately risky bike infrastructure ranges from 37% in Chicago to 11% in Las Vegas, to somewhat risky areas ranges from 20% in Louisville to 0.8% in San Jose, and to slightly risky areas ranges from 7% in New York to 0% in San Jose. Figure 5 represents the socially vulnerable populations exposed to extremely risky bike infrastructure for a 20 min commute. Each bar is the share of the population for a specific vulnerable group in extremely risky areas divided by the share of the population for the same vulnerable group in non-extremely risky areas. This ratio is deemed an inequity indicator with values greater than 1, meaning the disproportionate exposure of a vulnerable population to extremely risky bike infrastructure. The results reveal that African Americans and Hispanics are the most vulnerable groups facing inequity of access to well-connected low-stress bike infrastructure. Carless households also mostly reside in areas with a large BARG and are disproportionately exposed to extremely risky bike infrastructure for reaching employment opportunities. In addition to the individual assessment of the various vulnerable groups in figures 5(a)-(e), we collectively examine all groups in figure 5(f).

Social equity of bike infrastructure
For African Americans, as the most vulnerable group, we find that they disproportionately reside in areas with large BARG for 44 out of 50 of the metropolitan areas. In Milwaukee, approximately 2.5 times more African Americans reside in extremely risky areas compared to non-extremely risky areas. This is followed by roughly a similar share for Indianapolis and Hartford. African Americans living in San Diego, Pittsburg, and Buffalo have the least issues of BARG inequality. Hispanics are the most vulnerable in Milwaukee, Indianapolis, and Providence by a margin of 2.2 times, 2 times, and 1.8 times, respectively. Boston, Buffalo, and Cleveland have the least BARG inequalities affecting Hispanics. Low-income households do not experience the social inequity seen by African Americans or Hispanics, and there are also fewer metropolitan areas affected by the inequity of access. We find that low-income residents of Austin, Jacksonville, and Indianapolis experience the most BARG inequality, while Pittsburg, Riverside, and Portland residents are least affected by BARG inequality. Residents of Austin and Jacksonville experience social inequity of access due to their lack of cars by a margin of 1.6 times. The elderly is the most protected group as we show fewer metropolitan areas are affected by BARG inequality. Collectively, more metropolitan areas (76%) are affected by BARG inequality. Milwaukee has the largest margin of social inequity of the BARG.
We expand our analysis by performing the Pearson correlation between BARG calculated for a 20 min commute and for socially vulnerable populations as well as population density in each metropolitan area at the census block group level. The results are documented in supplementary data III. The association was found positive for the elderly in nine metropolitan areas, for African Americans in 44 metropolitan areas, for Hispanics in 36 metropolitan areas, for low-income households in 26 metropolitan areas, for carless households in 30 metropolitan areas, and for population density in 36 metropolitan areas. This indicates areas with higher population density, albeit offering high access to employment opportunities, contain less well-connected low-stress bike infrastructure, and their populations are exposed more to extremely risky bike infrastructure. In contrast, areas with lower population density offer a well-connected, low-stress bike infrastructure with low access to employment opportunities mimicking a community suitable for biking with a not bikeable network. The Eugene and Madison metropolitan areas are two examples detailed in supplementary data IV.
Between 2001 and 2009, African American and Hispanic populations experienced the fastest growth rate in cycling (The League of American Bicyclists 2013). The growth in the percentage of all trips by bike was 100% for African Americans and 50% for Hispanics. However, our study shows the inequality of bike infrastructure access for African Americans and Hispanics. The focus on investments in bike infrastructure are evidently not steered and aligned toward the affluent in society even though African Americans and Hispanics' biking rate to work is higher than whites, and the percentage of work commute by bike is more common for low-income households in comparison to high-income households (The League of American Bicyclists 2018). As a result of low income, households become carless and more reliant on biking as a commute mode of travel. In these minority communities, there are also fewer bike infrastructures and facilities to support their biking trips. Accidents and crashes are an add-on consequence of the infrastructure gap. Here, we find city planning is not representative of the largest portions of the population that bike for work commutes. It would not be farfetched to see that the lack of equity of fund allocation results in the vulnerable becoming more vulnerable. Equality of outcomes is simply not the goal in identifying the class issues. We simply posit that the allocation of bike infrastructure funds and urban form should be linked with equity in the planning and execution that gives attention to the needs of the poor, vulnerable, and vulnerable groups. According to Golub (2014), the social inequalities of access are rooted in unequal access to (i) participation in the planning processes, (ii) localized environmental burdens, and (iii) distribution of mobility benefits from transport investments. The inability of city planners to design to the needs of community members due to the lack of participation in community forums is no longer an excuse for failure to understand the needs of the vulnerable. Unequal environmental burdens are associated with the varying transport access, quality, and experiences of concern to the localized community. Distribution of mobility requires investments to ensure social equities of access for all groups.
The vulnerable are less privileged in choosing their place of residence and employment compared to other groups. The choice of living close to work where housing prices are high is nonexistent, and the bike infrastructure does not support movements for far distances to the central business districts where a majority of jobs are centered. Farther commutes to work under strenuous conditions and high-stress bike infrastructure is the option as city planning efforts often exclude the vulnerable, with more attention given to a group with better social standing. Planning for equity is, therefore, necessary to fix the social inequities that value the higher social class of people over the needs of the socially vulnerable. This removes the economic prosperity chokeholds that improve the quality of life.

Conclusions
The study outcomes addressed elements of access to employment opportunities, infrastructure availability, connectivity, risk, infrastructure coverage, and socioeconomic and demographic advantage that are interwoven in the fabric of sustainability and equity of bike infrastructure. We draw three conclusions by assessing the temporal disparity, spatial disparity, and social inequity of bike infrastructure between and within metropolitan areas.
First, the probability of exposure to extremely risky bike infrastructure increases with an increment in bike commute time. The number of employment opportunities reachable by exposure to extremely risky bike infrastructure becomes dramatically higher than the number of employment opportunities reachable by exposure to slightly risky bike infrastructure beyond a 20 min commute. Second, a large proportion of land is covered by a wide BARG. However, the proportion of distance of bike networks with low-level traffic stress is higher regardless of the metropolitan area (Owen and Murphey 2020). A wide BARG, which to a certain extent is due to the lack of well-connected, low-stress bike networks can be bridged by constructing low-stress bike infrastructure on or near high-stress bike infrastructure. Third, vulnerable populations are disproportionately exposed to extremely risky bike infrastructure for commuting. African Americans, Hispanics, low-income, and carless households experience a wide BARG. The role of equity in transport seeks to guarantee fair treatment in access, opportunity, and advancements for all groups of individuals while by the same token identifying and eliminating barriers. This evidence of bike infrastructure access favoring one group over another is a question of planning and investment. Better bicycle infrastructure is the result of better-funded neighborhoods and communities through infrastructure investments. The potential to narrow or close the inequity divide in biking infrastructure availability and accessibility requires infrastructure investments for better bike infrastructure planning.
The temporal disparity, spatial disparity, and social inequity of bike infrastructure is a policy product that might overlap with other dimensions of inequities (e.g., noise exposure, pollution exposure) and racism of urban planning (e.g., redlining, segregation) at the national or regional level. There is a need to better identify communities experiencing 'compound inequities' to adjust recent federal initiatives (e.g., Justice40) and arm policymakers and planners to strategically allocate resources to reduce the inequity gap between vulnerable and non-vulnerable populations.

Data availability statement
The data that support the findings of this study are openly available at the following URL/DOI: https://doi. org/10.13020/mdpv-ec74.