Innovations in Bus Rapid Transit: Event Recap


By Carson Qing

This week, the NYU Rudin Center and the Wagner Transportation Association (WTA) hosted a panel discussion of recent innovations in bus rapid transit (BRT) in the New York City metropolitan area. The panel’s presenters included Ted Orosz from MTA New York City Transit, Eric Beaton from the New York City Department of Transportation, and Tom Marchwinski from New Jersey Transit.

The discussion highlighted how transportation providers were able to find innovative solutions to implement BRT under the unique context of the New York City metropolitan region, where street widths, curbside usage, land use characteristics, and competing transit options often pose challenges for developing a BRT system similar to those built in Latin America and Asia. The panel’s speakers highlighted how implementation of Select Bus Service in New York City and bus rapid transit in high-volume, medium-density, and suburban settings in New Jersey have succeed in reducing travel times, improving level-of-service, and attracting new riders by adapting BRT characteristics to better fit the context of the corridors and communities they serve.

The presentations are available for download here: Ted Orosz, Eric Beaton, Tom Marchwinski

 

 

Super-Commuting vs. Mega-Commuting


Carson Qing & Sarah Kaufman

Earlier this week, The U.S. Census released a report announcing the proliferation of “mega-commuters,” 600,000 Americans who travel at least 90 minutes and 50 miles each way. It’s slightly different from the “super-commuters” we at the NYU Rudin Center defined last year, who are individuals who work in one county (usually of a major metropolitan area), but live in another, usually commuting more than 90 miles each way.

The most pressing difference between the terms “mega-commuter” and “super-commuter” is that the former focuses on the individuals traveling long distances regularly to their workplaces, while the latter also includes people who make these journeys once or twice or week, at most. These long-distance, low-frequency super-commuters may travel to the office only once or twice per week at most, or maintain similarly unconventional schedules. Our definition of a super-commuter, estimated to be 3% to 10% of the workforce depending on the city, includes both “mega-commuters” and low-frequency, long-distance commuters who were not captured in the mega-commuter definition. The graphic below illustrates the differences between these two types of super-commuters in their travel behavior.

 

The U.S. Census Bureau provides two data sources to define origins and destinations of commuter flows. To define the mega-commuter, the Census Bureau used American Community Survey (ACS), which measures data from only 7.5% of the working population, then extrapolates the data for a larger population based from that sample. But the Census Bureau’s OnTheMap tool (OTM), used in our super-commuter report last year, extracts employment data directly from state employment insurance records and represents coverage of nearly all employees and their work locations, with the exception of self-employed individuals. Because of this difference between ACS and OTM, the “mega-commuter” figure is most likely an undercount of long-distance commuters.

Using OTM, we found nearly 650,000 long-distance commuters in the top five U.S. super-commuting metropolitan areas who commute to the core county from a county outside the metropolitan area. OTM is more successful at capturing low-frequency commuting trips than the ACS, because the ACS’s line of questioning focuses on frequent trip-making, asking respondents where did they work for the majority of the past week and how did they travel to work, and assumes that the sample data applies to a larger population[1]. Low-frequency commuters are coded as “working from home” in the ACS, even though in reality their link to the workplace is not severed: the trips are made less frequently, due to the impediments of travel time, distance, and cost.

The rise of “tele-commuters,” who now represent 10% of the total workforce (or in the case of Aetna, 47% of its workforce, up from 9% in 2005[2]), and low-frequency, long-distance commuting has created a fundamental shift in the way people travel between home and work. The traditional “Journey to Work” survey methodology used in the ACS does not fully capture new patterns of commuting or the growing distances between home and work locations in metropolitan regions. It neglects the large and growing number of Americans who do not travel exclusively between home and work on a regular basis. Thus, transportation planners and researchers should not overly rely on the “Journey to Work” methodology to analyze and understand transportation flows: a more nuanced data source that captures a greater variety of trip purposes is increasingly necessary to analyze travel behavior in this new era of commuting.


[1] Spear, Bruce. “Improving Employment Data for Transportation Planning.” Cambridge Systematics. September 2011. http://onlinepubs.trb.org/onlinepubs/nchrp/docs/NCHRP08-36(98)_FR.pdf

[2] Miller, C. & Rampbell, C. “Yahoo Orders Home Workers Back to the Office.” The New York Times. 25 February 2013.

Manhattan Commuting Trends: An In-Depth Look


Carson Qing

Earlier this week, we discussed the unique patterns of employment “re-centralization” that the New York City metropolitan area experienced over the past decade. Now, we focus on the region’s core, Manhattan, and where its commuters are coming from. A detailed analysis, building on last year’s report describing trends in commuting among Manhattan’s workforce, reveals that most of the growth in Manhattan commuting has originated from waterfront neighborhoods in Jersey City, Hoboken, and Brooklyn, areas that experienced significant high-density residential development in recent years.

Using the Longitudinal Employer-Household Dynamics dataset from the U.S. Census Bureau, I identified specific towns and neighborhoods (defined as ZIP codes) that have the greatest increase in commuters to Manhattan. The interactive map below shows areas of residence with growth and declines in Manhattan commuters from 2002 to 2010 in absolute numbers. Zip codes shaded as blue represent a decrease or no difference in commuters to Manhattan. Darker shades of red indicate greater increases in commuters to Manhattan from that zip code. Click around to see the figures at a neighborhood level.

These numbers indicate substantial increases in Manhattan work trips originating from Northern Brooklyn, Western Queens, Jersey City and Hoboken, the South Bronx and Staten Island. The five neighborhoods with the greatest increase in Manhattan commuters were Williamsburg (+5,405), the Paulus Hook section of Jersey City (+4,262), Downtown Brooklyn (+3,598), Williamsburg/Bedford-Stuyvesant (+3,373), and Greenpoint (+3,139), all consisting of neighborhoods situated along either the Hudson or East River waterfronts. Areas that saw declines in commuters to Manhattan were largely in the northern and eastern suburbs, consisting of neighborhoods in eastern Queens and Westchester, Rockland, and Nassau counties.

High-density residential developments along the waterfronts in New Jersey, Brooklyn and Queens, paired with the expected increase in Manhattan-bound commutes from those neighborhoods, indicate that there are significant opportunities for expansion in ferry services in New York City. The East River Ferry that connects the neighborhoods of Downtown Brooklyn/DUMBO, Williamsburg, Greenpoint and Long Island City with the Midtown East and Lower Manhattan business districts has been far more successful than originally anticipated during the first year of its 3-year pilot service, carrying more than 1.6 million passengers (300,000 more than expected). A long-term extension and expansion of ferry services on the East River should be strongly considered as a strategy to relieve rush hour crowding on subway lines such as the L and 7 lines and provide a more convenient travel alternative.

The growth in Manhattan commuting to from the west in suburban New Jersey is not limited to communities with “one-seat” rides into Manhattan where no transfers are required to get in. Communities in Bergen and Passaic Counties along the Main-Bergen and the Pascack Valley rail lines, where Manhattan-bound rail trips require transfers at either Secaucus Junction or Hoboken to enter Manhattan, have also seen significant increases in commuters to Manhattan: these include towns such as Fair Lawn (+39% increase), Paramus (+30%), and Lodi (+47%). Workers traveling to Manhattan from those areas are much more dependent on the regional express bus system operated by NJ Transit and private companies to commute into Manhattan, and will continue to be dependent due to the cancellation of the Access to the Region’s Core (ARC) rail infrastructure project in 2010. Making the region’s system of commuter buses run more efficiently, whether by creating additional capacity at the Port Authority Bus Terminal or providing an express bus lane in the Lincoln Tunnel during evening rush hour, should help accommodate this growth in commuters from suburban New Jersey and sustain the region’s economic productivity and competitiveness in the 21st century.

 

News at the Rudin Center


The NYU Rudin Center staff has been busy:

Rudin Center Director Mitchell Moss discussed the making of Hipsturbia and organic dry cleaners as indicators of gentrification in The New York Times.

Research Associate Sarah Kaufman will present the Rudin Center’s report on Superstorm Sandy at the Transportation Equity Conference in Albany on March 4th.

Research Assistant Carson Qing‘s study of Williamsburg’s late night rush hour has been featured in the Brooklyn Paper and The L Magazine. His newest post on location of employment in major U.S. is now on the blog.

We’re proud to bring on Anthony Townsend as Senior Research Fellow. Here’s a look at the work he’ll be doing at the Rudin Center:

Anthony Townsend is organizing several upcoming workshops that will further the Rudin Center’s investigations into emerging areas of transportation policy, planning and management – resilient regional transportation infrastructure for the Northeast Corridor, future tools and techniques for studying bicycle ownership and use in New York City, the role of big data and pedestrians, and future mobility systems in digitally-connected cities. Through his affiliation with the Silicon Valley-based Institute for the Future, Anthony is conducting a year-long forecast on the future of makers and small-scale manufacturing in cities around the world. His first book, SMART CITIES: Big Data, Civic Hackers and the Quest for a New Utopia will be published in October 2013 by W.W. Norton & Co.

 

Finally, some of our research staff attended the State of the City address at Barclays Center. Here’s a photo:

 

Be sure to follow us on Twitter and Facebook for regular updates.

The State of Employment Decentralization in Major American Cities


Carson Qing

Since the mid-20th century, employers have followed its employees to the suburbs, and have adapted the workplace to fit their employees’ commuting needs, leading to the rise of the “corporate park” and the “edge city.” Some scholars have observed that in the 2000s, a dramatic shift has occurred as cities were again attracting the jobs that left in earlier decades, as employers respond to changing preferences among younger workers who desire a more urban lifestyle. Others contend that such a conclusion is premature, and that employment decentralization, also known as “job sprawl,” still occurs, as there is still high demand for suburban living. Using data on private sector employment from the Census Bureau’s Local Employment Dynamics, I tried to determine if the pattern of employment distribution across metropolitan areas had truly shifted in the past decade, and based on my findings, it seems that job distribution and movement vary by region, although generally, the trends remain slightly in favor of continued employment decentralization in major U.S. metro regions.

Metro regions with an increase in the share of its workforce employed clustered within 5 miles of the Central Business District were:

  1. San Francisco: +1.5%
  2. New York: +1.3%
  3. Detroit: +0.8%
  4. Chicago +0.2%
  5. Philadelphia +0.1%

The above cities are all older designs, where most development occurred early in the 20th century, in the pre-automobile era. Metro regions with the greatest increase in the share of its workforce employed within 20 – 50 miles of the CBD (or, “job sprawl” tendencies), were:

  1. Atlanta: +4.5%
  2. Dallas: +2.9%
  3. Houston: +2.6%

These cities are generally sprawling, Sun Belt areas that have experienced much of its growth during the late 20th century. After accounting for job trends based on distances from each region’s CBD, I observed the following patterns of employment growth (see methodology below for more detail):

Jobs in New York and San Francisco are increasingly concentrated in their urban core. In these cities, employment is no longer de-centralizing, but is re-centralizing. Both cities have a dense and diverse urban core that offer distinctive amenities and advantages for workers and employers, which could be a major driver of these recent trends.

A group of cities had an increasing share of jobs in both its urban core and its exurban fringes, but a smaller share in the “core-periphery” area: the peripheral areas of the primary city, and inner-ring suburbs that border the city. These cities exhibit a “U-shaped” relationship between the increase in the share of jobs in a given zone and the distance from the center city. One-third of the metro areas sampled exhibited this spatial pattern of job growth, including Chicago, Philadelphia, Atlanta, Detroit, and St. Louis.

In Houston and Dallas, employment decentralization has been sustained. Areas further from the city are capturing a greater share of the region’s jobs. This trend resembles the traditional pattern of late-20th century employment decentralization.

In general, employment decentralization has been sustained in the largest metro regions in the United States since 2002, but mostly at the expense of the “in-between” zones situated within 5 to 10 miles of the CBD, rather than the CBD itself. These generalized job growth trends show that the past decade was a period of deepening spatial divisions within U.S. cities. Overall, diverging demographic preferences and market forces are leading to an unconventional pattern of employment distribution, one that places the high-density urban core and the low-density suburban fringes at a distinct advantage over the medium-density urban periphery and inner-ring suburbs, locations that typically do not offer the agglomeration advantages of the central city, nor the accessibility advantages of the exurban fringes.

 

Methodology:

This analysis divided the 15 largest metro regions (defined as all census tracts within 50 miles of the primary city’s CBD) into 4 zones of analysis, based on distance from the city center. After calculating job growth for each of the zones and for each metro region, the data was smoothed to reflect a “best-fit” trendline. A composite average of the job growth data was also obtained and fitted to a trendline (highlighted by the red curve above). The composite average trend indicates that regional trends generally favor sustained employment decentralization, but there are distinctive variations across metro regions and the spatial patterns are more complex than anticipated.

The fitted trendlines of New York and San Francisco are negatively sloped (highlighted in yellow), which indicates that recent job growth and distance from the city center appear to be inversely related and have a highly linear pattern.

The fitted trendlines of Houston and Dallas are positively sloped (highlighted in blue), indicating that areas further from the city are capturing a greater share of the region’s jobs.

Rush Hour in Williamsburg…at 1 AM


By Carson Qing

Last September, one of our research assistants at the NYU Rudin Center, Nolan Levenson, took an interesting picture at the Bedford Avenue subway station in Williamsburg, Brooklyn (right). The subway platform was filled to capacity with straphangers, but what makes the photo interesting is that the image was captured in the wee hours of a Sunday morning, at 1:30 AM. There has been much discussion, and subsequent action, over the issue of providing more L-train service on the weekends to better serve this ridership growth, but the image of a subway platform filled to near capacity at 1:30 AM on a Sunday morning, when Manhattan-bound trains run on 20 minute headways, raises some interesting questions about travel characteristics along this particular subway line.
Since 2005, ridership on the L train has soared, with every station in Brooklyn posting double digit growth rates in ridership on weekdays (with the exception of Broadway Junction). On weekends, ridership by station has grown at even faster rates: tripling or even quadrupling the ridership growth on an average weekday for a given station. The Morgan Avenue station in Bushwick had the greatest ridership growth on both weekdays (+59%) and weekends (+174%) of all L-train stops in Brooklyn from 2005 to 2010. The Bedford Avenue station in Williamsburg had the greatest absolute increase in average weekday ridership (+5,867) and average Saturday ridership (+9,236) from 2005 to 2010. The two maps below compare ridership growth on an average weekday (left) and on average weekend (right) for all L-train stations in Brooklyn, from 2005 to 2010.

 

To examine these weekend ridership trends in more detail, I used the MTA’s turnstile data and took a sample of a turnstile at the Bedford Avenue station over one week in August 2012 to identify trends in peak hours of subway ridership, and what could be driving these patterns in weekend ridership. I classified both entries and exits into the Bedford Avenue station and identified “peak hours” in subway ridership, which were hourly intervals that were in the top 25% of all intervals in the sample data in total entries or exits into the station. The results are summarized in the chart below (note: data is only for a single turnstile, and is only meant to illustrate ridership trends):

What’s remarkable about this case study for Bedford Avenue is that not only are there ridership peaks for long durations on Saturday (8 am to 4 am Sunday) and Sunday (8 am to 8 pm), but entry/exit figures are actually comparable to morning and evening rush hours during the work week: thus, growth in weekend ridership at Bedford Avenue has increased so much that it may very well have resulted in an “extended rush hour” for almost the entire weekend.

Even more remarkable is that the peak entry hours on Saturday night actually extend into the wee hours of Sunday morning for the sampled data, suggesting that the crowded subway platform at 1:30 AM might in fact be quite a common occurrence. Given recent, dramatic changes in demographics and land use patterns in Williamsburg, these unusual peak hour trip patterns should be expected. Not only has there been a well-documented influx in 25-to-34 year olds in Williamsburg (25% of the population, compared to 17% in 2006, according to census data), but there has also been a significant growth in restaurants and bars that are open late on weekends and draw young New Yorkers from across the city to the neighborhood (117% increase in full service restaurants and 59% increase in bars since 2005, according to census business data). The peak entry hours from 12 am to 4 am on a Sunday morning should be expected given the context of how Williamsburg has changed dramatically in just a few short years, as many of the restaurant and bar patrons are likely contributing to this peak period of subway ridership during these late night hours.

These trends reveal that due to the growth in weekend ridership on the L-train, conventional assumptions of travel demand for this particular subway line may no longer be appropriate, and may require some adjustments in service offerings during weekend evenings, late nights, and other times of day. According to subway schedules, the MTA currently runs roughly 43 Manhattan-bound trains on the L during a weekday morning rush hour (8 am-12 pm) and 48 Manhattan-bound trains during Saturday afternoon (4 pm-8 pm), falling to roughly 32 on Saturday night (8 pm -12 am) and 13 during weekend late-night hours (12am-4 am Sunday). With only 13 trains during one of the busiest travel periods of the entire week, crowded platforms at Bedford Avenue and nearby stations during late Saturday nights/early Sunday mornings will likely be commonplace going forward.

The growth in weekend ridership on the L-train in Brooklyn and peak travel demand during unconventional hours show how as cities and neighborhoods evolve, traditional assumptions of “rush hour” travel will inevitably change. Transportation providers should be flexible and adaptable to recognize these anomalies, rather than assume that travel characteristics are uniform system-wide, and respond by offering level of services that are appropriate given these unique patterns in peak travel demand.

Have you taken the L from Bedford Avenue during late night hours on the weekend? Are weekend, late night hours in Williamsburg comparable to weekday morning “rush hours?” Please share your experiences in the comments below.

Commuting After Hurricane Sandy: Survey Results


Sarah Kaufman and Carson Qing

As part of the NYU Rudin Center’s recent report on transportation impacts from Hurricane Sandy, we conducted a survey of commuters to learn about their experiences of getting to work after the storm.

The survey was conducted online, on the site Surveymonkey.com, and was publicized via email blasts and social media. Three hundred-fifteen people in 98 zip codes responded anonymously between October 31 and November 6th, answering questions about their typical and post-Sandy commutes.

Key findings from the survey included:

Many people in the region worked after the storm, either physically reporting to an office or by telecommuting. New Jersey had the lowest rate of people who continued to work, at 56%, while 85% of Brooklyn respondents worked, at the highest percentage.

With limited transit options after the storm, New York commuters significantly altered their commute patterns. Bus ridership rose in Brooklyn (5% of respondents normally used buses, but 12% reported using buses November 1-2) after shuttle buses were put in place of subway routes disrupted due to flooding. Bike commuting rose significantly in Manhattan (15% normal to 24% Nov 1/2) and Queens (17% to 30%).

Post-hurricane commute lengths varied significantly by home region, as shown in the table below. The largest differences were in Staten Island, where commute times almost tripled, and Brooklyn, where they doubled. Variations among home locations are due to the wide range of transportation options available to each set of commuters, and the lower number of survey respondents who reported physically to work, rather than telecommuting or not working.

Post-hurricane commutes were twice or three times as long, varying by mode, as shown in the chart below.Average post-Sandy commute lengths ranged from 43 minutes (walked on Nov 1/2) to 115 minutes (drove, or took subway and bus). Frustration levels ranged from 2.3 on the lower end (walked) to 5.7 on the higher end (drove). Commuters who drove, or took a subway and bus combination, had the greatest difference, with travel times at nearly triple their typical lengths. As expected, they were also among the most frustrated commuters.

Walking and biking commuters were, on average, the least frustrated. Commuters who biked to work Nov 1/2 had the fewest delays in their commutes, as they were only 9 minutes longer than their usual commute. Telecommuters ranked their level of frustration on a similar level as transit commuters, 3.7 to 3.8, perhaps due to communications difficulties of connecting to work.

Commuters used a variety of communications channels to learn about transportation resources, as shown in the chart below. They most commonly referred to official websites and social media, and least from smartphone apps and community groups. The lack of smartphone app connectivity was likely due to the lack of schedule and outage data used for programming the apps.

These figures show the need for increased storm preparation and ever-present public information in times of crisis to ensure residents’ mobility. However, the survey results also demonstrate the resilience of New Yorkers and their workplaces; even in the face of detrimental circumstances, New Yorkers’ businesses maintained operations, showcasing the extreme adaptability of their operations, finances and creativity. The adaptations to new, longer commutes are uniquely New York, in that the population quickly adapted to alternate and substitute transportation modes, new norms of local business practices, flexible, temporary workplaces, and continuous communications.

 

Survey respondents’ home and workplace locations, by zip code:

 

 

Average commute times and frustration levels by home region, November 1-2, 2012

Home Region Pre-Sandy Typical Commute Time (min) Post-Sandy Commute Time (min) Percent Reporting Physically to Work* Self-Reported Frustration Level, 1 (min) – 10 (max)
Manhattan 29 52 56% 2.97
Brooklyn 42 86 58% 3.93
Queens 45 47 65% 3.00
Bronx 41 63 100% 2.14
Staten Island 84 240 25% 7.00
New Jersey 52 69 27% 5.67
Northern Suburbs 73 61 33% 2.40
Long Island 85 85 33% 2.00

* Excludes telecommuters

 

 

Commuters’ travel time by mode and self-reported frustration level:

NOV 1/2 MODE Pre-Sandytravel time (min) Post-Sandy travel time (min) Avg frustration index (1-10)
Walk only 21.1 43.3 2.3
Bike only 43.6 52.0 2.7
Drive only 47.3 114.7 5.7
Taxi only 30.0 65.0 5.5
Subway only 35.0 51.4 2.9
Bus only 42.3 100.8 4.2
Rail only 80.0 85.0 2.0
Subway + bus 46.5 115.1 4.9
Subway + bus + rail 60.0 75.0 2.0
Any transit* 41.7 86.3 3.8
Telecommuting 40.1 0.0 3.7
Did not work 42.3 0.0 5.6

*includes PATH, private buses, ferries and other miscellaneous transit options

 

Sources of Transportation Information

Respondents were asked to select all that apply.