Replicable Methodology for Transportation Agencies to Identify Priority Areas for First and Last Mile Solutions at the Regional Level

Abstract

First and last mile (FLM) solutions that enhance connections to fixed-route transit offer multifaceted benefits, yet relatively few transportation agencies have developed comprehensive regional strategies or plans to support implementation. There is a lack of guidance and replicable methodologies that transportation agencies can use to conduct regional spatial analysis or screening for priority areas to invest in FLM solutions at a regional scale. This paper presents a replicable, step-by-step methodology and accompanying visual aids for metropolitan planning organizations (MPOs), transportation management associations (TMAs), and transit agencies to identify priority areas for investment in FLM solutions. This methodology was developed through two related studies conducted in 2021 by the North Jersey Transportation Planning Authority (NJTPA), the MPO for the 13-county northern New Jersey region. This methodology can be replicated using readily available data from the U.S. Census Bureau and transit agencies, along with simple analytic processes in geographic information systems (GIS) software including buffers, spatial difference, intersection, and filtering. Agencies can tailor this methodology based on the types of fixed-route transit facilities under consideration, the distances that FLM solutions may cover, and whether agencies are seeking to prioritize employment clusters, residential clusters, and/or transit hubs for future investment. The paper concludes with a discussion of the various applications and limitations of regional analyses and planning for FLM solutions.

Keywords

planning and analysis congestion mitigation or reduction planning methods regional planning public transportation GIS accessibility and mobility

First and last mile (FLM) solutions are transportation services and infrastructure that enhance connections to and from fixed-route transit: the “first mile” from a trip origin to transit, the “last mile” from transit to the trip destination, or both. These solutions can have an outsized impact on the total demand for travel, reducing vehicle miles traveled (VMT), while also improving regional accessibility by filling gaps in the multimodal transportation network ( 1 ). In fact, changing the FLM mode of access to transit can be more effective at improving regional accessibility than transit service enhancements such as reducing wait times or improving headways ( 2 ). Opportunities to enhance mobility and accessibility through FLM solutions are particularly salient in the wake of the devastating and sustained impacts of the COVID-19 pandemic on transit ridership, resulting in widespread reductions in transit service frequency and coverage ( 3 , 4 ).

Despite the multifaceted benefits of implementing FLM solutions, relatively few transportation agencies have developed comprehensive regional strategies or plans at a regional scale related to FLM implementation ( 5 – 7 ). In this paper, a regional scale is defined as spanning multiple counties and including a mix of urban, suburban, and rural settings. Research on existing FLM planning efforts has highlighted spatial gap analysis (focusing on socio-demographics and locational characteristics) as a primary planning approach, along with incorporating emerging mobility services, innovative funding approaches for plan implementation, and equity and transportation remedies for marginalized communities, and developing pedestrian and bicycle infrastructure surrounding transit stations ( 8 ). Among these approaches, spatial gap analysis at the regional level should be a valuable first step for transit agencies, metropolitan planning organizations (MPOs), and transportation management associations (TMAs) to identify priority areas for further study and investment in FLM solutions. Without such analysis or regional screening, the selection of FLM investments may be overly influenced by local interests or opportunistic factors without sufficient consideration of public needs and regional goals.

There is a lack of easily replicable methodologies that transportation agencies can use to conduct regional spatial analysis or screening for FLM priority areas without proprietary or highly technical tools or datasets. Existing research and methodologies include an environmental audit methodology for assessing the quality of FLM connections to transit, incorporating walkability, pedestrian level of service, and feeder bus metrics, which focuses more on characteristics of the built environment than on land use and accessibility to jobs or households ( 9 ). Another study explores data and analysis tools for using accessibility measures and GPS-based trip-making data to inform potential FLM solutions ( 10 ). The author notes that accessibility estimates require network data and the ability to calculate travel times between all origins and destinations in a system, and thus require sophisticated tools and data inputs. The methodological complexity to generate an accessibility measure, such as the number of jobs accessible within a 30 min transit ride at every point across a region, poses a barrier for transportation agencies with limited resources. Other researchers have developed a methodological framework for analyzing the FLM opportunities in relation to public transit accessibility using only publicly available open datasets, including General Transit Feed Specification (GTFS) and Census Transportation Planning Products (CTPP) ( 11 ). However, this framework requires implementing an algorithm that is beyond the technical capacity of most transportation planning practitioners. This lack of easily replicable approaches for regional planning of FLM reflects the wide variation in the availability of relevant data, tools, and technical capacity for this analysis at transit agencies, counties, municipalities, MPOs, and TMAs across the U.S.

The objective of this paper is to present a replicable methodology for transit agencies, MPOs, and TMAs to identify priority areas for investment in FLM solutions using transit hub locations and/or location of employment and residential clusters. This methodology can be applied using readily available data (including data from the U.S. Census Bureau and data on transit facility locations) and simple analytic processes in geographic information systems (GIS) software including buffers, spatial difference, intersection, and filtering. This methodology was developed through two related studies conducted in 2021 by the North Jersey Transportation Planning Authority (NJTPA), the MPO for the 13-county northern New Jersey region.

Study Background

In 2021, NJTPA completed a regional Transportation Demand Management (TDM) and Mobility Plan as well as the Accessibility and Mobility Strategy Synthesis, an update to the federally required regional Congestion Management Process (CMP) ( 12 , 13 ). Both of these planning initiatives identified FLM solutions as a priority strategy for advancing regional goals related to accessibility, mobility, and reducing VMT. The study team conducted regional analyses of priority areas for FLM solutions for both studies to serve as a replicable model to inform further study and implementation by NJTPA and other partners. The next section describes the approaches to each of these analyses in greater detail.

Methods

The study team implemented two analysis methods, one each for the TDM and Mobility Plan and the CMP update, which NJTPA developed concurrently in 2021. From these two analyses, the study team created a replicable methodology, presented at the beginning of this section. The TDM and Mobility Plan analysis identified clusters of workers and jobs within 2–5 mi of both rail stations and park and ride lots. The results identified activity centers needing longer-distance connections to transit through shuttles, transportation network company (TNC) partnerships, and carpooling. Concurrently, the CMP update identified FLM transportation as a congestion management strategy and presented a regional analysis of access needs at transit hubs, based on rail stations with an above-median concentration of workers and jobs within 0.5–2 mi, as well as park and ride lots with over 100% utilization. The results of this analysis identified transit hubs that should be prioritized for short-distance FLM solutions including infrastructure and services supporting active transportation and micromobility.

Together, these analyses represent priority areas at both ends of an FLM trip: the work or home location as the origin or destination, and the transit hub through which travelers connect to fixed-route transit service. While NJTPA conducted these two analyses separately, they coordinated the methods and results. Both analyses excluded areas within 0.5 mi of bus stops because bus service could presumably serve as a feeder or FLM service to rail or express bus routes, or walking could be viable. However, the analysis did not evaluate the suitability of individual bus routes by examining destinations, frequency, and service day. Such further study would be a component of a site analysis.

This section presents the data sources used for both analyses, then the generalized methodology (merging the two approaches) that agencies can replicate and tailor to suit local transportation characteristics, available data, and priorities for investing in FLM solutions. This section concludes with a summary of the methods for each study, with the TDM and Mobility Plan prioritizing residential and employment clusters and the CMP update prioritizing transit hubs.

Data Sources

Data sources and types used for the analyses are outlined below:

Census blocks boundaries for the regional study area

Longitudinal Employer Household Dynamics (LEHD) Origin Destination Employment Statistics (LODES) attributes joined to census blocks:

○ number of jobs per census block (used to identify employment clusters)

○ number of workers per census block (used to identify residential clusters, based on workers’ home locations)

Transit agency data:

○ rail station locations

○ bus stop locations

○ park and ride lot locations

Additionally, the CMP update analysis used park and ride lot utilization rate data to identify lots that were oversubscribed and would benefit from FLM solution or additional parking capacity. Park and ride lot utilization rate data may not be available or applicable in regions where park and ride lots are rarely used at capacity.

Replicable Methodology for Regional Analysis of FLM Priority Areas

The study team combined the two regional analyses for prioritizing FLM solutions described above to create a replicable methodology that regional transportation agencies can implement using readily available data from the U.S. Census Bureau LODES dataset and geolocation of transit facilities (e.g., rail stations, bus stops, park and ride lots). The analysis steps use simple tools in GIS software including buffers, spatial difference, intersection, and filtering.

While the NJTPA analyses focused on FLM solutions to rail stations and/or park and ride lots, this methodology is generalized so as to be mode-agnostic. Thus, the methodology can be applied to any transit facility, including rail, park and ride lots, and highly serviced bus stops (high frequency or express) for regions that lack rail or wish to enhance FLM connectivity to bus service. Additionally, this methodology is region-agnostic in that it can be applied across urban, suburban, and rural regions of all sizes across the U.S.

Table 1 below outlines the replicable 10-step methodology for prioritizing employment and residential clusters (in the left-hand column) and for prioritizing transit hubs (in the right-hand column) for investment in FLM solutions. Steps that are the same for both approaches are merged into a single column. For ease of understanding, the reader is recommended to read one column at a time. As previously discussed, the approaches together represent priority areas at both ends of an FLM trip: the work or home location as the origin or destination, and the transit hub through which travelers connect to fixed-route transit service. Agencies can choose to conduct one or both of these analyses. It is recommended that agencies start with the left column in Table 1 to identify employment and residential clusters before prioritizing transit hubs, because land use and trip generators (rather than transit facilities) are the primary drivers of travel demand. However, transit agencies focusing on increasing ridership or use of transit facilities may choose to first conduct the analysis prioritizing transit hubs. Agencies may also choose to tailor the distance range beyond the two options presented (0.5–2 mi and 2–5 mi) based on local context. Table 1 is followed by a list of additional, optional steps through which agencies may refine the results to account for worker and job characteristics, connecting transit services, park and ride lot utilization rates, and directionality of transit service.

Table 1.

Replicable Methodology for Regional Analysis of First and Last Mile (FLM) Priority Areas

General process	Detailed steps
General process	Prioritizing employment and residential clusters	Prioritizing transit hubs
Setting up the base data layer	Step 1: Set up a geographic information systems (GIS) layer or multiple layers with all transit facilities to consider for FLM solutions including rail stations, highly serviced bus stops (with high frequency or express service), and/or park and ride lots in the region. This is the base data layer.
Selecting the distance range to transit for analysis	Step 2: Select the range of distance to/from transit stops that potential FLM solutions would cover. This distance would guide interpretation of which FLM modes would be best suited to priority areas and may be based on regional needs, goals, or objectives. • 0.5–2 mi: FLM modes that are best suited to trip distances in this range include active transportation and micromobility options. • 2–5 mi: FLM modes that are best suited to trip distances in this range include shuttles, transportation network companies (TNCs), and carpooling.
Setting up the target area to be analyzed	Step 3: Create buffers around the base layer for low and high end of distance range (e.g., 0.5 mi and 2 mi, or 2 mi and 5 mi). Subtract the lower buffer from the higher buffer using the “difference” tool in GIS to create donut-shaped polygons.
	Step 4: Create a 0.5 mi buffer around the bus stops layer and select. If the base data layer includes highly serviced bus stops, then those should be excluded from this step, so that only local or feeder bus service is buffered.	Step 4: Identify transit hubs with high number of connecting bus routes (e.g., three bus routes). These transit hubs are considered to have a relatively high level of existing connectivity.
	Step 5: Select the donut-shaped polygons created in Step 3 around the base layer, then subtract the 0.5 mi buffers of bus stops created in Step 4. This is the target area.	Step 5: Remove the transit hubs identified in Step 4 from the base data layer. Remove the polygon buffers associated with those points as well. The remaining polygons comprise the target area.
Analyzing worker population in the target area	Step 6: Filter census blocks within the target area with high worker population (e.g., the top 25th percentile of the target area. Blocks with over 100 workers were used in the North Jersey Transportation Planning Authority [NJTPA] rail analysis.)	Step 6: For each donut-shaped polygon within the target area, sum the number of workers within all census blocks that intersect (even partially). This sum represents the approximate number of workers within 0.5–2 mi of each transit hub.
Analyzing worker population in the target area	Step 7: To further refine target areas, filter census blocks within the target area with the highest worker population using a higher threshold than in Step 6 (e.g., the top 15th percentile of the target area. Blocks with over 500 workers were used in the NJTPA rail analysis.)	Step 7: Filter stations ranked in the top 50th percentile of the number of workers between 0.5 and 2 mi of each transit hub.
Analyzing number of jobs in the target area	Step 8: Filter census blocks in the target area with a high number of jobs (e.g., the top 25th percentile of the target area. Blocks with over 200 jobs were used in the NJTPA rail analysis.)	Step 8: For each donut-shaped polygon within the target area, sum the number of jobs within all census blocks that intersect (even partially). This sum represents the approximate number of jobs within 0.5–2 mi of each transit hub.
Analyzing number of jobs in the target area	Step 9: To further refine target areas, filter census blocks in the target area with the highest number of jobs using a higher threshold than in Step 8 (e.g., the top 15th percentile of the target area. Blocks with over 1,000 jobs were used in the NJTPA rail analysis.)	Step 9: Filter transit hubs ranked in the top 50th percentile of the number of jobs between 0.5 mi and 2 mi of each transit hub.
Generating final results	Step 10: Show census blocks in the target area with the highest number of workers and jobs. They represent the locations that will benefit most from FLM solutions.	Step 10: Show transit hubs displayed in Steps 7 and 9. These transit hubs represent the locations that will benefit most from FLM solutions.

Results of the analysis can be refined further through the following optional steps:

LODES data on jobs and workers per census block can be filtered to only include those defined as low-income (with earnings of $1,250 per month or less). This ensures that the regional analysis of potential areas to invest in FLM solutions prioritizes low-income workers and jobs, providing an equity lens to the analysis. Aside from earnings, LODES data on jobs and workers can also be filtered by industry in the North American Industry Classification System (NAICS), worker age, gender, educational attainment, race, firm age, and firm size ( 14 ).

To better understand regional commute flows, planners may wish to investigate the home locations of workers affiliated with employment clusters, and, vice versa, the job locations of workers affiliated with residential clusters. To visualize commute flows to, from, and within specific residential and/or employment clusters, use the Census OnTheMap web tool to conduct inflow/outflow analysis ( 15 ). The inflow/outflow analysis will provide the number of workers that: (1) are employed in the cluster but living outside it; (2) are living within the cluster but working outside it; and (3) both live and work within the cluster ( 16 ). Note that the highest level of spatial granularity available through this tool is the census block group. As a result, the inflow/outflow analysis will not correspond exactly with the census blocks used in the FLM analysis to identify residential and employment clusters.

Eliminate rail stations that have many connecting bus services (identified as three or more unique buses within a 0.5 mi of the station). Stations with many bus services in the area could be considered as generally having a high level of connectivity, and thus one may wish to prioritize other stations for new or additional services, particularly in areas that may not have the density of transit ridership to provide additional bus services. The resulting stations can be identified as potentially promising to consider for FLM connection improvements.

If the park and ride lot utilization rate data are available, filter lots to prioritize those that are under- or over-used. The CMP update analysis identified park and ride facilities that are oversubscribed (over 100% utilization, based on data pre-COVID-19) for possible access improvements (which could be in the form or micromobility, shuttles, or other services) or for stations in more rural and suburban areas that might be considered for parking expansion.

Identify transit hubs that serve only express or interstate bus or rail routes to evaluate suitability of FLM solutions. For these express or interstate transit hubs, only first-mile connections from workers’ homes to the transit hub should be considered. These transit hubs likely will not benefit from last-mile connections to jobs, since they provide long-distance service to destinations (such as New York City in the North Jersey region).

Overlay worker and household transportation characteristics at the census block level, such as vehicle ownership and commute mode share. This information would help prioritize target areas based on the goals of an FLM program, for example, to improve mobility options in an area with many zero-car households, or to shift drive-alone commuters to transit in a congested area.

To further refine priority areas based on equity considerations, and overlay additional datasets displaying environmental justice indices, communities of concern, or other measures of vulnerable populations.

As a visual aid to accompany Table 1, Figure 1 below illustrates a sample sequence of key steps in GIS for prioritizing employment clusters located 2–5 mi from rail stations. These images were developed through the NJTPA TDM and Mobility Plan FLM analysis.

Figure 1.

Sample sequence of key steps in geographic information systems (GIS) for prioritizing employment clusters for first and last mile (FLM) solutions.

Prioritizing Employment and Residential Clusters

The TDM and Mobility Plan analysis focused on prioritizing employment and residential clusters at the census block level that were located between 2 mi and 5 mi from rail stations and park and ride lots (measured in Euclidian distance). The catchment area of 2–5 mi was selected based on a mode-based typology table of FLM solutions developed for the TDM and Mobility Plan, which identified shuttles, TNCs, and carpooling as best suited to this distance range. While the CMP update analysis focused on priorities at the “transit” end of an FLM trip, this analysis focused on the other end: the residential and employment activity centers that need to connect to transit.

Analysis of priority employment and residential clusters for FLM solutions were conducted separately for rail stations and park and ride lots (serving bus transit); however, they involved very similar methods. The TDM analysis for employment and residential clusters around rail and park and ride lots is outlined below with respect to the steps of the generalized methodology provided in Table 1.

Step 3: Donut-shaped polygons were created outside a 2 mi buffer of a rail station or park and ride lot (serving bus transit) but within a 5 mi buffer. They were created by subtracting the 2 mi buffers from the 5 mi buffers using the difference function in GIS to create the polygons.

Step 4: Buffers of 0.5 mi were created around all bus stops in the region. The 0.5 mi bus stop buffers were subtracted from the donut-shaped polygons to create the target area for analysis.

Step 5: The resulting area is considered as the “target” area for conducting the FLM needs analysis. It represents areas that are close enough to a transit hub to facilitate an FLM solution, such that they do not have access to bus transit which could support the FLM travel.

Steps 6 to 10: Within the target area, census blocks were identified with high worker population and/or number of jobs. This analysis displayed census blocks with “high” worker/job population and the “highest” worker/job population. The “high” and “highest” thresholds were selected based on the distribution of population and employment in the NJTPA region. The thresholds for worker population were 100 workers per census block (“high”) and 500 workers per census block (“highest”). The thresholds for job population were 200 jobs per census block (“high”) and 1,000 jobs per census block (“highest”). Generally, the “high” threshold can be around the top 25th percentile and the “highest” threshold can be around the top 15th percentile.

Optional step: Park and ride lots were differentiated that serve only interstate or express bus commuter service. For these interstate/express-only park and ride lots, only first-mile connections from workers’ homes to the park and ride lots should be considered. These park and ride lots likely will not benefit from last-mile connections to jobs, since they serve peak-direction express bus service to interstate destinations such as New York City or other employment centers.

Prioritizing Transit Hubs

The CMP update analysis focused on prioritizing transit hubs, specifically rail stations that have an above-median density of residential or job locations within 0.5–2 mi (measured in Euclidian distance) as well as oversubscribed park and ride lots. For purposes of this analysis, a geographic area of between 0.5 mi and 2 mi around each rail station was selected as a reasonable distance where connections via bicycling, scooters, or other micromobility options, as well as localized shuttles or coordination with private providers, could help to enhance access to transit. The selection of the 0.5–2 mi catchment area was informed by federal guidance on improving bicycle and pedestrian connections to transit ( 17 ). The analysis included park and ride lots where parking demand was very high compared with the number of available spaces, as those lots may benefit from shuttles or other access connections, or possibly additional parking capacity.

The regional FLM analysis used in the NJTPA CMP update is outlined below with respect to the steps in the generalized methodology provided in Table 1.

Step 3: Donut-shaped polygons were created outside a 0.5 mi buffer of a rail station but within a 2 mi buffer. They were created by subtracting the 0.5 mi buffers from the 2 mi buffers using the difference function in GIS to create the polygons.

Steps 4 and 5: Stations and associated donut-shaped polygons were eliminated that have many connecting bus services (identified as three or more unique buses within 0.5 mi of the station). Stations with many bus services in the area were considered as generally having a high level of connectivity, and thus one may wish to prioritize other stations for new or additional services, particularly in areas that may not have the density of transit ridership to support additional bus services. The resulting stations were identified as potentially promising to consider for FLM connection improvements.

Steps 6 and 7: Rail stations were identified that had a sufficient density of workers within 0.5 mi and 2 mi of each station. Specifically, rail stations that met any of the criteria below were considered potential priorities for FLM solutions based on the density of potential users:

○ stations ranked in the top 50th percentile of the number of workers between 0.5 mi and 2 mi of each rail station

○ stations ranked in the top 50th percentile of the number of low-income workers between 0.5 mi and 2 mi of each rail station

Steps 8 and 9: Rail stations were identified that had a sufficient density of jobs within 0.5 mi and 2 mi of each station. Specifically, rail stations that met any of the criteria below were considered potential priorities for FLM solutions based on the density of potential users:

○ stations ranked in the top 50th percentile of the number of jobs between 0.5 mi and 2 mi of each rail station

○ stations ranked in the top 50th percentile of the number of low-income jobs between 0.5 mi and 2 mi of each rail station

Additional step: In addition to the stations identified above, park and ride facilities that are oversubscribed (over 100% utilization, based on data pre-COVID-19) were also identified for possible access improvements (which could be in the form of micromobility, shuttles, or other services, or, for stations in more rural and suburban areas, consideration of parking expansion).

Results

This section presents the results of the regional FLM analyses conducted through the TDM and Mobility Plan (prioritizing employment and residential clusters) and the NJTPA CMP update (prioritizing transit hubs). These results illustrate what the outputs of these analyses may look like to MPOs, TMAs, and transit agencies considering whether to apply in their own regions the approaches described in the Methods section.

Prioritizing Employment and Residential Clusters

The results of the regional FLM analyses conducted in the NJTPA TDM and Mobility Plan identified priority employment and residential clusters. These clusters represent census blocks with the highest numbers of workers (>500) and jobs (>1,000) within the regions that were located 2–5 mi from rail stations and park and ride lots.

In the maps that follow, census blocks with the highest numbers of workers are shown in dark blue, while census blocks with the highest numbers of jobs are shown in red. The few Census blocks that contain highest numbers of both workers and jobs are displayed as purple, where blue and red polygons overlap.Figure 2 shows the census blocks with the highest numbers of jobs and workers located 2–5 mi from rail stations, and Figure 3 shows the census blocks with the highest number of jobs and workers located 2–5 mi from park and ride lots (serving bus transit). As discussed in the Methods section, park and ride lots serving only interstate bus routes were symbolized differently, since these locations will likely not benefit from last-mile connections to jobs but may still benefit from first-mile connections to workers’ residences.

Figure 2.

Regional results from first and last mile (FLM) analysis for census blocks within 2–5 mi of rail stations.

Figure 3.

Regional results from first and last mile (FLM) analysis for census blocks within 2–5 mi of park and ride lots serving bus corridors.

The mapped results are followed by Table 2, an excerpt from a regional summary table that shows the number of priority census blocks located within each county and municipality. This table accompanies the maps as a quick-reference and alternative format for regional and municipal planners to interpret results of the analysis and identify municipalities that may have a high need for FLM solutions.

Table 2.

Excerpt from Summary of Priority Areas by Municipality

County	Municipality	Number of priority census blocks per municipality
		Rail first and last mile (FLM) analysis		Park and ride FLM analysis
		Workers > 500	Jobs > 1,000	Workers > 500	Jobs > 1,000
Hudson	Jersey City	0	0	1	2
	Bayonne	0	0	0	1
	Secaucus	0	0	1	0
Hunterdon	Branchburg Township	0	1	0	0
	Franklin Township	1	0	0	0
	Readington Township	0	1	0	0
	Union Township	1	0	0	0
Middlesex	Piscataway Township	1	12	4	6
	Edison Township	0	2	2	7
	Woodbridge Township	0	2	3	5

Prioritizing Transit Hubs

The results of the regional FLM analyses conducted in the NJTPA CMP update identified priority transit hubs (rail stations and park and ride lots) as shown in Figure 4 below. The priority rail stations represent stations with above-median concentrations of workers and/or jobs within 0.5–2 mi and are shown in red, while the priority park and ride lots are utilized at over 100% of capacity and are shown in green. Rail stations with above-median concentrations of jobs and workers that also have oversubscribed park and ride lots are shown in blue.

Figure 4.

Priority transit hubs for first and last mile (FLM) solutions as a result of the Congestion Management Process (CMP) update analysis.

Discussion

Transportation agencies and partners can use the results of these regional analyses and/or refine the methodologies to determine areas of opportunity and high need for enhanced FLM services. For example, in regions where housing prices are significantly higher near transit stops, demand for FLM services among lower-income and transit-dependent populations may extend well beyond the distance ranges recommended in this analysis. Once priority census blocks with high numbers of workers and/or jobs are identified, partners can conduct a local area analysis at the scale of a county or municipality to better understand local conditions and identify potential locations for site-level evaluation. The local area analysis should include industry classifications, equity considerations such as neighborhood demographics, and transit characteristics such as service capacity, directionality, and reverse peak direction demand to employment destinations such as warehousing and distribution centers. Planners may need to consider the impacts and aftermath of the COVID-19 pandemic on operations and ridership of existing fixed-route bus and employer shuttles, along with resulting uncertainties about future demand and service levels. The local area analysis should also be easily replicable and scalable across geographic areas with data updated as demographics, transportation networks, and population and employment centers shift in the future.

Once priority areas are identified, it will be important to match needs with applicable FLM solutions through a coordinated planning, design, and implementation process. Following demand analysis, design and delivery of FLM investments may involve microsimulations of services and/or environmental review, engineering, and construction of infrastructure ( 18 ). To facilitate further analysis and application of FLM solutions in North Jersey, the NJTPA TDM and Mobility Plan team developed an implementation brief that identifies potential actions along with partners, resources, next steps, and timeframes ( 19 ). The brief creates a general framework for FLM implementation by combining the regional analysis described in this paper with a descriptive typology of FLM solutions by mode, as well as a site-level decision tool. Together, these materials inform preliminary FLM design decisions with respect to distance range, place type of origin or destination, service area, supportive infrastructure and safety, operational models, partnership configurations, payment mechanisms, funding models, and environmental impacts. In the months following completion of the TDM and Mobility Plan and CMP update, NJTPA has coordinated with federal and regional transit partners to develop a future FLM grant program based on these analyses.

Limitations

These regional analyses attempted to identify potentially promising locations to consider for strategy implementation based on readily available data within the CMP update and TDM and Mobility Plan development processes. However, a wide variety of other factors, including existing conditions in the corridors, public and stakeholder input, and local interests, should play an important role in making decisions about investment in FLM solutions. Thus, it is important to note that the regional analyses presented in this paper are insufficient on their own to justify public investment. For example, this analysis only considered absolute numbers of jobs and workers to identify priority employment and residential clusters, which may bias the results toward geographically larger census blocks. Further analysis accounting for employment and residential densities may be needed to control for this bias. Additionally, this analysis was limited by its use of Euclidian distance (as the crow flies) rather than network distance (based on street configurations) for the 0.5 mi, 2 mi, and 5 mi buffers around transit facilities. Euclidian buffers are simple to execute in most GIS applications, but they tend to overestimate access sheds because FLM travel is typically limited to street networks. While GIS applications differ in their network analysis capabilities, network-based buffers can depict access to transit with higher accuracy, helping planners to refine priority areas for FLM investment. Because of these limitations, this methodology should be considered a preliminary analysis for identifying areas for possible consideration.

More specific studies—whether region-wide, corridor, or sub-area based—will need to be conducted to assess issues such as transit service needs, development potential, roadway geometrics and operations, travel demands, and qualitative factors. Additional datasets may address: pedestrian/bicycle level of comfort (based on presence, quality, and connectivity of bicycle and sidewalk networks and roadway functional class); transit supply (GTFS data can provide extent of transit coverage based on time of day and week, and, specifically, feeder bus catchment areas could be informed by speeds, frequencies, and fares); and location-based services data that reflect travel patterns and mode choice. Within the North Jersey region, many studies have already been undertaken or are ongoing to help identify and assess various strategies. This analysis could be a starting point to consider and compare areas identified through the CMP update and the TDM and Mobility Plan as potential areas of promise for FLM solutions.

It should also be noted that FLM solutions do not work in all contexts. In rural and low-density areas without transit service, a complete-trip solution provided through a ridematching system, TNC partnership, or demand-response shuttle may be needed to transport users between their origin and destination rather than to or from a transit stop.

Conclusions

Implementing FLM solutions has the potential to advance regional goals related to mobility and accessibility, yet there is a lack of guidance and methodologies that transportation agencies can use to identify priority areas to invest in FLM solutions based on regional needs. This paper provides a high-level and replicable methodology for screening priority areas for FLM solutions based on two analyses conducted through the NJTPA CMP update and TDM and Mobility Plan. MPOs, TMAs, and transit agencies considering a regional strategy or planning framework to extend the reach of fixed-route transit can tailor this methodology using readily available data and simple analytics processes in GIS. Agencies can prioritize employment clusters, residential clusters, and/or transit hubs based on regional priorities. Additional analysis at region-wide, corridor, or sub-area-based scales could assess issues such as transit service needs, development potential, roadway operations, travel demands, and qualitative factors. For example, an in-depth review may include fixed-route transit coverage and ridership by day and time of day as well as cost-benefit analysis. Additionally, agencies could overlay activity on transportation networks beyond the fixed-route transit system such as bicycle and pedestrian counts, traffic volumes, parking demand, and location of trips by micromobility and/or TNCs to prioritize transit hubs or activity centers. The results of regional analyses supported by this methodology provide a foundation for more rigorous investigation at the local and/or site-level of potential areas that may be well-suited for public investment in FLM solutions.

Footnotes

Author Contributions

The authors confirm contribution to the paper as follows: study conception and design: K. O’Sullivan, D. Shah, P. Bilton, E. McGuiness; data collection: D. Shah; analysis and interpretation of results: K. O’Sullivan, D. Shah, P. Bilton, E. McGuiness; draft manuscript preparation: K. O’Sullivan. All authors reviewed the results and approved the final version of the manuscript.

Declaration of Conflicting Interests

The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.

Funding

The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The development of the methodology and the analysis were sponsored by the North Jersey Transportation Planning Authority.

ORCID iDs

Katie O’Sullivan

Eugene McGuinness

References

Shaheen

Chan

Mobility and the Sharing Economy: Potential to Facilitate the First- and Last-Mile Public Transit Connections. Built Environment, Vol. 42, No. 4, 2016, pp. 573–588.

Boarnet

M. G.

Giuliano

Hou

Shin

E. J.

First/Last Mile Transit Access as an Equity Planning Issue. Transportation Research Part A: Policy and Practice, Vol. 103, 2017, pp. 296–310. https://doi.org/10.1016/j.tra.2017.06.011.

Liu

Miller

H. J.

Scheff

The Impacts of COVID-19 Pandemic on Public Transit Demand in the United States. PLoS One, Vol. 15, No. 11, 2020, p. e0242476.

Bellis

What Service Cuts are Transit Agencies Facing Around the Country?

T4 America. December 2020. https://t4america.org/2020/12/01/what-service-cuts-are-transit-agencies-facing-around-the-country/.

Utah Transit Authority. First/Last Mile Strategies Study. April 2015. https://www.rideuta.com/-/media/Files/About-UTA/Tiger-VIII/UTAFirst_LastMileFINALCOMP1.ashx?la=en.

Regional Transportation District. First and Last Mile Strategic Plan. 2019. https://www.rtd-denver.com/sites/default/files/files/2019-07/FLM-Strategic-Plan_06-10-19.pdf.

Los Angeles County Metropolitan Transportation Authority. First Last Mile Strategic Plan & Planning Guidelines. March 2014. https://media.metro.net/docs/First_Last_Mile_Strategic_Plan.pdf.

Mohiuddin

Planning for the First and Last Mile: A Review of Practices at Selected Transit Agencies in the United States. Sustainability, Vol. 13, No. 4, 2021, p. 2222. https://doi.org/10.3390/su13042222.

Venter

C. J.

Measuring the Quality of the First/Last Mile Connection to Public Transport. Research in Transportation Economics, Vol. 83, 2020, P. 100949. https://doi.org/10.1016/j.retrec.2020.100949.

10.

McCahill

Improving Last-Mile Connections to Transit: An Exploration of Data and Analysis Tools. Proc., 25th Annual Congress for the New Urbanism. May 2017. https://www.cnu.org/sites/default/files/2017_NewUrbanResearch_ImprovingLastMileConnectionstoTransit_McCahill.pdf.

11.

Liu

X. C.

Porter

R. J.

Zlatkovic

Fayyaz

Taylor

First and Last Mile Assessment for Transit Systems. MPC-18-347. North Dakota State University - Upper Great Plains Transportation Institute, Fargo: Mountain-Plains Consortium, 2018.

12.

North Jersey Transportation Planning Authority. Transportation Demand Management and Mobility Plan. June 2021. https://www.njtpa.org/Planning/Regional-Programs/Studies/Completed/2021/Transportation-Demand-Management-and-Mobility-Plan.aspx.

13.

North Jersey Transportation Planning Authority. Accessibility and Mobility Strategy Synthesis. June 2021. https://www.njtpa.org/Planning/Regional-Programs/Congestion-Management/Accessibility-Synthesis.aspx.

14.

U.S. Census Bureau. LEHD Origin-Destination Employment Statistics (LODES) Dataset Structure Format Version 7.5. November 2020. https://lehd.ces.census.gov/data/lodes/LODES7/LODESTechDoc7.5.pdf.

15.

U.S. Census Bureau. OnTheMap. N.d. https://onthemap.ces.census.gov/.

16.

U.S. Census Bureau. OnTheMap Sample Inflow/Outflow Analysis. December 2011. https://lehd.ces.census.gov/doc/help/onthemap/OnTheMapSampleIOAnalysis.pdf.

17.

McNeil

Dill

DeVitis

Doubleday

Duncan

Weigand

Manual on Pedestrian and Bicycle Connections to Transit. Federal Transit Administration Report, FTA Report No. 0111. August 2017. https://www.transit.dot.gov/sites/fta.dot.gov/files/docs/research-innovation/64496/ftareportno0111.pdf.

18.

Los Angeles County Metropolitan Transportation Authority. First/Last Mile Guidelines. May 2021. https://www.dropbox.com/s/395zn5ghsi1uqvy/Metro-FLM-Guidelines.pdf?dl=0.

19.

North Jersey Transportation Planning Authority. First/Last Mile Solutions. Implementation Brief. June 2021. https://www.njtpa.org/NJTPA/media/Documents/Planning/Regional-Programs/Studies/Transportation%20Demand%20Management%20and%20Mobility%20Stud/First-Last-Mile-Implementation-Brief.pdf?ext=.pdf.