Office of Operations
21st Century Operations Using 21st Century Technologies

City of San Francisco, California: Advanced Transportation & Congestion Management Technologies Deployment Initiative (ATCMTD)- 2017

Section I: Project Narrative

What is the Next Generation Transit Customer Information System?

The San Francisco Municipal Transportation Agency (SFMTA) is developing a new real-time vehicle arrival and service update system for the Muni public transportation network which it operates. Known formally as the Next Generation Transit Customer Information System, the project is designed to increase public confidence in Muni so that customers can take transit to their destinations quickly and reliably. By retaining and growing transit ridership, the system would contribute to the city and region's mobility, accessibility and economic health.

As the United States looks to an increasingly urbanized future, public transportation will become more important than ever to the nation's economic well-being. By attracting more ridership to space-efficient and environmentally-friendly transit, this Next Generation Transit Customer Information System is essential to building a more sustainable San Francisco. With support from the US Department of Transportation's Advanced Transportation and Congestion Management Technologies (ATCMTD) program, the innovations developed under this project will provide a blueprint for other cities looking to strengthen their transit systems.

Guided by public feedback, the system will employ the latest technology to make transit easier and more convenient to ride in an environment of increasing transportation choices.

Examples of innovations include:

Developing a more sophisticated and accurate vehicle arrival prediction algorithm
Communicating service delays and disruptions on-board vehicles
Implementing solar-powered signage to expand access to information at selected unpowered shelters and stops
Strengthening network connectivity by showing transfer connection times on-board vehicles
Reducing travel times by showing nearby alternative routes with shorter waits on dig- ital signage at stops
Providing stop accessibility information and elevator/escalator outage alerts for seniors and people with disabilities
Balancing capacity by providing crowding alerts and suggesting parallel services with space available
Using data from mobile technologies to understand customer preferences and improve service and operational planning


Figure 1 : San Francisco's Projected Growth by 2040

Figure 1: San Francisco's Projected Growth by 2040 (Source: San Francisco Planning Department)

The Context

Public transportation in the United States is at a crossroads. More than 50 years after the federal government officially became involved in funding transit, national ridership has surpassed 10 billion annual trips. Ridership is up by nearly 60 percent since the all-time low in 1972, led primarily by more than a doubling of ridership on rail modes as dozens of American cities have opened or expanded their heavy rail, commuter rail and light rail systems.

Recently, however, transit has struggled to hold onto these gains. In 2016, for example, national ridership fell by 2.3 percent from the previous year. More alarmingly, ridership on buses - which operate primarily in mixed traffic along with low-occupancy automobiles and ride-sharing vehicles - plunged 4.1 percent at a time when the economy and transportation demand grew. This is not sustainable. With space at a premium, America's metropolitan areas increasingly must depend on high-capacity public transportation in order to accommodate long-term population and employment growth and remain economically strong.

Nowhere is this fact more evident than in San Francisco, the second-densest major city in the United States that is constrained on three sides by water and on the remaining side by mountains. Between 2010 and 2016 alone, U.S. Census Bureau data indicate that the residential population increased from 805,235 to 870,887 (8.1 percent growth). According to the Bureau of Labor Statistics, City employment surged from 544,963 to 701,520 (26.4 percent growth) during the same period. Looking into the future, the San Francisco Planning Department projects that the city could have 100,000 new households and 190,000 new jobs by 2040. With limited space for road capacity and parking, there is simply not enough room to design a transportation system centered around low-occupancy vehicles.

Rather, San Francisco and the Bay Area's livability and economic prosperity will depend on more people walking, biking and riding transit.

Many reasons have been posited for recent national ridership declines such as low gas prices, the economic recovery and the emergence of on-demand ride-sharing services. While these may be key factors, SFMTA's research suggests that innovations in real-time transit customer information will be essential to reverse these trends and retain and grow transit ridership.

More than 15 years ago, the SFMTA became a national pioneer in real-time transit information when it implemented its first generation system on its Muni transit network. Before then, people typically would venture out to a stop randomly and just wait until a bus or train arrived, however long it took, because published timetables were not reliable. The first generation system fundamentally changed how people used transit by giving them an expectation of when Muni would come once they arrived at a stop. This basic model, largely still employed by transit systems in the United States, delivers vehicle arrival predictions through signs at stops and now through mobile apps and other means. It has become a vital tool for time management when customers are preparing to ride a bus or rail vehicle or even deciding what form of transportation to take in the first place.

Times have changed, however. San Franciscans, like other Americans, have grown accustomed to a vastly different technological landscape that continues to redefine the transportation industry through on-demand service offerings and mobile information before and during one's trip.

As SFMTA's contract with the first generation vendor will soon expire, we are looking to leverage the many innovations in technology and transportation that have occurred in the past 15 years to create a much more advanced Customer Information System that satisfies and even exceeds today's customer expectations. As a multi-modal transportation organization, the SFMTA adheres to the City's Transit-First Policy, which prioritizes transit, bicycling and walking. While it recognizes that people may use other modes such as private vehicles and ride-sharing services, the agency also operates one of the most comprehensive transit networks in the United States such that traveling by car should be an option, not a necessity within San Francisco. By making the extensive Muni network easier to use, the SFMTA believes that the Next Generation Transit Customer Information System can help San Francisco realize substantial latent transit demand.

Expanding real-time arrival signage, displaying transit alternatives during long waits, providing crowding alerts, offering real-time transfer connection times on-board vehicles and keeping our customers continually informed of service delays and reroutes throughout their journey are examples of innovations implementable through existing technology. These will make Muni and complementary sustainable modes like walking and bicycling more attractive travel choices.

That is why we are seeking a partnership with the U.S. Department of Transportation through the Advanced Transportation and Congestion Management Technologies (ATCMTD) program to help turn these ideas into reality. To our knowledge, many of the features under consideration have yet to be implemented elsewhere in the United States. We believe that our Next Generation Transit Customer Information System will produce practical, scalable and repeatable solutions that can other cities can deploy to improve their transit systems.

Why Focus on Real-Time Customer Information?

The experience waiting for and riding public transportation is a significant determinant as to whether one chooses it over other options. Long waits, overcrowding, not knowing when a bus or train will arrive, recurrent delays without explanation and missed transfers can easily drive customers to less sustainable forms of transportation. In contrast, shorter waits, real-time arrival information, minimal delays and timely transfer connections provide a show-up-and-go convenience and a positive experience that encourages more discretionary trips and customer loyalty.

Service Reliability

Real-time information throughout one's journey is essential because customers value service reliability. Transit systems that operate in mixed traffic, the vast majority of service in the United States, have challenges maintaining scheduled service frequency. Actual service differs from schedules because of traffic congestion, variable passenger loads, the deployment of wheelchair ramps, and other factors. All Muni routes, including the Muni Metro light rail system, operate all or portions of their services on surface streets, making them likely to encounter headway variability. For example, a 15-minute route might actually have two vehicles spaced 10 minutes apart, followed by a 20 minute gap.

Service Frequency

Service frequency is one of the most important factors in a person's decision on whether to take transit or another form of transportation. Unfortunately, transit systems in the United States usually lack sufficient operating funds to provide as much as service as they would like. Anything they can do to lower actual or perceived waiting times - including customer time management through real-time information - can directly contribute to ridership retention or growth.

Muni offers some of the highest levels of transit service in the country, making it possible to travel anywhere within San Francisco on transit. Not only is nearly everyone within three blocks of a Muni route, but most bus and rail routes operate from 6 a.m. to 11 p.m. or later every day and usually at least every 15 minutes during peak periods. However, there are also weaker parts of the network. In some of the outer neighborhoods or in general during the late evenings, most routes operate every 20 to 30 minutes.

Figure 2: Muni Service Frequency

Rush Hour Service. General Streep map of San Francisco. Most routes are green indicating less congestion and less reduced speeds.

Rush Hour Service
Generally every 15 minutes or more often.

Late Evening Service.  A General Street map of San Francisco.  Only a few green sections.  Most are streets are colored orange or red indicating heavy congestion and reduced speeds.

Late Evening Service
Generally every 20 - 30 minutes

Even with this high level of service, customer expectations of maximum waiting times are also high. As indicated in Figure 3, without knowing any real-time arrival information, most people are willing to wait at least 10 minutes and perhaps up to 15 minutes. Few people are willing to wait 20 minutes or more, which is how often many routes are operating at night of other off-peak periods. There is also a noticeable decline in willingness to wait for transfers.

Figure 3: Muni Service Frequency

Figure 2: Muni Service Frequency

While Muni's frequent service during the daytime aligns fairly well with customer expectations, the discrepancy between their willingness to wait and the frequency of service suggests the importance of having real-time information particularly during off-peak periods and for transfers.

Project Vision, Goals and Objectives

The overarching goal of the Next Generation Transit Customer Information System is to increase public confidence in Muni so that customers can take transit to their destinations quickly and reliably. Specific project goals include:

Provide accurate real-time information
Increase discretionary and off-peak ridership
Offer alternatives during long waits or service delays
Understand how information and service quality impact customer travel choices
Retain customers who might otherwise use less sustainable transportation modes
Strengthen relationship with Muni customers

In addition, this project will improve:

Connectivity

This project will use real time arrival predictions to facilitate intra- and inter-agency transfers to other transit operators such as BART, provided that those systems use a standard software interface, and enable customers to take advantage of the interconnected transit network within San Francisco and to the rest of the region.

Reliability

While the new system would not improve transit reliability per se, it would generate real-time transit alternatives in case of long waits, service disruptions, or delays. By enabling customers to take full advantage of the density and comprehensiveness of the Muni network, the new system would reduce their waiting time and thus their actual travel time, which would make their Muni experience more reliable and less stressful.

Access

This project will provide San Francisco residents and visitors with greater access to destinations in San Francisco by making transit options clearer and easier to use. Customers will have a better understanding of what services are available in real-time, which would help them navigate a complex system and lower barriers to transit usage.

Equity

This project prioritizes providing information to people without smartphones or data plans. It will devote significant resources to expand digital signage to select, currently unpowered shelters and stops without shelters, including in low-income and minority communities.

Additionally, the Next Generation Transit Customer Information System fulfills the following specific goals of the Advanced Transportation and Congestion Management Technologies (ATCMTD) program.

Reduced costs and improved return on investments, including through the enhanced use of existing transportation capacity

The system would inform potential customers to nearby Muni services when they might have otherwise chosen less efficient transportation modes. It would also provide crowding alerts and suggest nearby less-crowded alternative routes if available; some people could alter their transit itinerary such that loads are more evenly spread across routes. This would allow the SFMTA to improve use of existing transit capacity and accommodate customers more optimally and comfortably using existing resources. By smoothing out demand, vehicles will have less crowding and less wear and tear.

Demonstration, quantification, and evaluation of the impact of these advanced technologies, strategies, and applications

To the SFMTA's knowledge, the Next Generation Transit Customer Information will incorporate many enhancements to real-time information that have not been widely implemented elsewhere in the United States. It will also generate a large amount of data, including by the system itself and system users, that can be matched with ridership and operational data to understand the impacts of an advanced real-time customer information system. SFMTA's customer research to date indicates a high potential for project to change user behavior. The project includes a robust analytics component as well as independently-conducted data interpretation to quantify impacts on customer satisfaction, mode choice and ridership.

Delivery of environmental benefits that alleviate congestion and streamline traffic

By making it easier to ride Muni, this project will shift people from less environmentally-sustainable modes to transit. Muni generates about 90% less greenhouse gas emissions and other air pollutants per trip (or mile) relative to automobiles, including ride-share vehicles. Muni's fleet - which includes zero-emission electric trolley coaches, light rail vehicles, historic streetcars and cable cars - is one of the greenest in the country.

Collection, dissemination and use of real-time transportation-related information to improve mobility, reduce congestion, and provide for more efficient and accessible transportation, including access to safe, reliable, and affordable connections to employment, education, healthcare, freight facilities, and other services.

Real-time transit information at the right times and places will make it easier to ride Muni, which will help increase discretionary ridership. It will also empower both regular and occasional customers to feel comfortable riding Muni to new or unfamiliar places, which will allow people to use transit to access multiple destinations including social services, jobs, businesses and medical facilities. As such, the project will shift people from less sustainable and less space-efficient modes to transit. A full Muni articulated (60-foot) bus displaces more than 90 single-occupant vehicles. A full two-car light rail train displaces over 235 single occupant vehicles. Shifting people from low-capacity to high-capacity vehicles will mitigate traffic congestion on San Francisco's limited roadway space.

Delivery of economic benefits by reducing delays, improving system performance and throughput, and providing for the efficient and reliable movement of people, goods and services

The SFMTA estimates that faster travel times from reduced congestion attributable to Muni ranges from $192.4-$236.8 million per year. A modest 3% ridership increase due to the project would yield $5.8-$7.1 million in monetized economic benefits alone. Transit is often less expensive than driving alone or using a ride-share service, which allows people to save more.

Reduction in the number and severity of traffic crashes and an increase in driver,passenger and pedestrian safety

By attracting riders to Muni, the project will improve overall system safety. On a per trip basis, the number of injuries and fatalities is about 77% on Muni than in automobiles, including ride-sharing vehicles. Fewer single-occupant vehicles, particularly in the downtown area and major corridors, would reduce the chances of automobile collisions with pedestrians and bicyclists to help the City achieve its Vision Zero safety goals.

Integration of Advanced Technologies into Transportation System Management and Operations

This project will take advantage of maturing and advanced technologies, including:

  • Sophisticated real-time transit arrival prediction algorithms
  • Algorithms that generate alternatives based on location and nearby vehicles
  • Mobile technologies and associated user information to understand system usage and customer preferences
  • Real-time passenger counting to assess vehicle loads
  • Real-time signage on-board vehicles
  • Solar-powered information signs

The project will employ advanced but readily available technologies that have been used in other industries but generally on a much more limited basis in public transportation.

Reproducibility of successful systems and services for technology and knowledge transfer to other locations facing similar challenges

Many cities are facing similar challenges as San Francisco with regards to long-term ridership trends, new transportation technologies and the potential for real-time information to impact them. The Next Generation Transit Customer Information System is large scale primarily due to the size and complexity of San Francisco's transit network and fleet. However, the project's elements are scalable to any city or region and have widespread applicability particularly for smaller and less extensive transit systems where service is less frequent and it can be more challenging to attract ridership that prioritizes short wait times. By engaging vendors and developing innovative system requirements, this project will help define best practices for real-time information systems for peer agencies and cities nationally.

Customer Research: Information at the Right Times and Places

To achieve key goals of the ATCMTD Initiative, the Next Generation project will need to increase Muni ridership. Why does the SFMTA believe that there is a high potential for this project will grow transit ridership? We asked our customers.

Informed by product development user research best practices, the SFMTA has embarked on an extensive customer research initiative to determine what is most critical to making informed travel decisions and incorporate them into system requirements. To date, public outreach has consisted of two primary ongoing efforts:

list

Quantitative

A broad-based online survey to capture the public transportation travel patterns, attitudes and preferences of as many Muni customers as possible. In addition to demographic and general travel behavior questions, the online survey contains hypothetical situational questions to assess the potential for providing transit customer information at the right times and places to influence travel behavior.

Group of People Image

Qualitative

To ensure that the SFMTA hears the opinions of Muni's diverse customer base, including voices that might be underrepresented in the online survey, the SFMTA is conducting focus groups to various community stakeholders including Senior and Disability Action, Independent Living Resource Center, LightHouse for the Blind and Visually Impaired and the Youth Commission.

On a scale of 1 to 5 (1 = poor, 3 = good, 5 = excellent), customers rated the overall quality of the current real-time information system a 2.5. This leaves ample room for improvement. High-level findings to date reveal that real-time information at the right times and places is essential to transit's usability and provides an opportunity for ridership growth.

How the Project Would Address Deterrents to Riding Transit

SFMTA's online survey asked customers to recall the last time they chose another form of transportation over Muni and to select up to two factors that influenced their decision. As indicated in Figure 4, the vast majority were service-related. Most importantly, the Next Generation Transit Customer Information System project has the potential to mitigate some of these factors and turn negative experiences into neutral or even positive ones. Over the long term, improved customer satisfaction should translate into ridership retention and growth.

Figure 4: How the Project Would Address Deterrents to Transit Ridership

Deterrent to Riding Transit Percent What the Next Generation Transit Customer Information System Will Do
Ride on-board Muni would take too long 34% While the new system would not directly address vehicle speeds, it could potentially recommend faster route options if available and reduce total travel time by shortening customer waits.
Scheduled service was too infrequent 33% While the new system would not directly increase service, it could reduce the negative impacts of infrequent service by allowing customers to delay their arrival at a stop or take a nearby alternative route coming sooner.
Muni did not arrive when predicted 30% By providing more accurate predictions, the new system would improve customer perceptions that Muni is "on time" and lower the chances  that customers would "give up" on Muni while waiting.
Muni was too crowded 22% While the new system does not directly increase resources to reduce overcrowding, it would provide crowding alerts to allow bus customers to wait for a later trip or find a nearby less crowded route.

Factor Percent What the Next Generation Transit Customer Information System Will Do
There was a service delay 22% While the new system would not directly eliminate service delays, it would inform customers of service disruptions or reroutes during their journey so they can seek alternatives and/or inform those waiting on them at their destination that they will be late.
Transfers were required 8% By providing real-time transfer connection times online and onboard vehicles, the new system would make transferring more predictable and give customers confidence they can take multiple Muni routes to more destinations.
I did not feel safe or secure 8% By providing arrival predictions that allow customers to reduce waiting times, the new system may improve perceptions of safety and security, at least while waiting for Muni.
Muni did not stop for me 6% A primary reason that vehicles do not stop for customers is that they often are full. By providing crowding alerts, the new system would allow bus customers to wait for a later trip or find a nearby less crowded route.

Source:SFMTA Next Generation Transit Customer Information System Survey(2,500 respondents as of June 7, 2017)

How Better Transit Customer Information Can Influence Mode Choice

As part of our public outreach, the SFMTA presented survey respondents with some situational questions to determine how different types of customer information delivered at various times could influence travel choices. As shown in Figure 5, we asked survey takers to imagine a situation where they were going home from work or school and had to wait 20 minutes for their Muni route. A 20-minute wait is not uncommon if there is a service gap or if one just missed a bus during the evenings or on weekends. Based on SFMTA's survey, only 15% of people indicated they are willing to wait this long for Muni.

Figure 5: Customer Information at the Right Times and Places Can Increase Ridership

Figure : Customer Information at the Right Times and Places Can Increase Ridership

In the base case, respondents arrive randomly at a stop and see a digital sign saying their route would arrive in 20 minutes. We asked respondents what they did the last time they encountered a situation. Then we tested a couple of scenarios. In one scenario, the real-time arrival sign at the transit shelter showed an alternate route three blocks away that was arriving sooner. In another scenario, they saw that their wait would be 20 minutes on their smartphone prior to walking to their stop.

The results were dramatic. In the base case, only 44% took Muni, either waiting the entire 20 minutes or finding an alternative transit route on their own. Displaying a nearby, earlier-arriving transit alternative on the real-time arrival sign at the stop nearly doubled the percentage of people who chose Muni to 82%. Additionally, 72% of respondents chose Muni when they could see the expected waiting time on a mobile device before starting to walk to the stop.

A final set of questions asked customers how they would travel if their trip required transferring between two Muni vehicles, with and without a prediction of the connection time. Assuming a hypothetical 6-minute wait, providing transfer time pre- dictions boosted the percentage of respondents would take Muni for at least a portion of their trip from 75% to 90% and for the entire length of their trip from 48% to 83%. Though responses to a stated preference survey do not always reflect revealed preferences, these numbers imply there is significant potential for the Next Generation Transit Customer Information System to increase Muni ridership.

Qualitative Research Findings

To date, the SFMTA has carried out 8 focus groups, 12 ride-along sessions, and 17 one-on-one interviews with a wide variety of Bay Area residents to gain a deeper understanding of the Muni customer experience. Our research has generated many findings that will assist in shaping the project.

For example, many customers understand there is traffic congestion and do not expect Muni to operate precisely according to the official schedule. However, they will consider Muni to be "late" if it does not arrive according to the real-time prediction displayed at a stop or on a mobile device. Therefore, by improving prediction accuracy, the SFMTA can also improve customer perceptions that Muni is reliable and operating "on-time."

Information Tools - customers are heavily reliant on technology for trip planning, including live maps.  Many customers simultaneously use multiple apps.  Many seniors and customers with disabilities prefer speaking with a live person on 311.  Accessibility - many customers with disabilities use Muni extensively and know routes well, but must monitor disparate sources of information to find out about accessible stops and elevator/escalator outages.

The SFMTA project team is currently organizing, coding and analyzing the qualitative data
collected from the first round of research to uncover more customer needs. In addition, the SFMTA will begin its next round of research by speaking with stakeholders in the San Francisco business and tourism community. To ensure the validity of research findings, the SFMTA will triangulate the qualitative data with the results from the on-going quantitative survey.

Branding - Muni does not have to be 'cool' like newer forms of transportation.  It has to function effectively within its constraints.  Perceptions of Time and Accuracy - Many customers perceive that a vehicle is 'late' when it does not arrive according to NextBus predictions.  This contrasts with the official definition of late (4 minutes later than the schedule).  Knowing the precise timetable is less than valuable than knowing one can arrive generally on-time.

Office of Operations