For the MTA, it’s BusTime vs. Countdown Clocks


As we know, New York City buses are slow and unreliable. Delayed by the vagaries of surface traffic, the city’s buses rarely stay on schedule and inch along surface streets. Buses are underutilized and often looked down upon by even their own riders. Few advances in the way we treat buses represent lost opportunities to move New Yorkers quickly and efficiently.

Over the past few years, as bus-tracking technology has swept the globe, New York has slowly embraced it. An expensive pilot program along 34th St. brought proprietary technology and countdown clocks to the heavily-trafficked corridor, but when the MTA searched for a system-wide solution, the agency instead went with a distance-based tracking system and no countdown clocks. BusTime is an open-source solution with flexibility for growth and real cost savings over closed systems.

BusTime, of course, isn’t perfect. It requires the user to possess a smart phone or texting capabilities and actually know that the technology is in place. It is a distance-based system, and for a bus to travel 1.3 miles depends upon the route, the time of day and the traffic in front of it. New Yorkers like to know time in minutes, not miles. (For a succinct summary of the MTA’s failed countdown clock efforts, check out this Daily News overview.)

Now, though, some politicians and advocates aren’t happy. Noting how countdown clocks make subway waits more tolerable and empower riders, they want countdown clocks, and they wait them now. Earlier this week, Council Member Brad Lander announced a new initiative aimed at convincing the MTA to install countdown clocks on bus shelters. Lander is leading the way with a piece of legislation that “calls upon the Metropolitan Transportation Authority, the Department of Transportation, and Cemusa to work together to install bus arrival time displays in bus shelters using data from the MTA’s Bus Time technology.”

Streetsblog had more from the press conference:

With countdown clocks already available in many subway stations, Lander and advocates say bus riders deserve the same convenience, and that not everyone has access to a cell phone or the Internet before catching a bus…Lander’s office estimates that the counters cost between $4,000 and $6,000 to purchase and between $1,000 and $1,600 to maintain each year, based on figures from other cities with bus countdown clocks, including Washington, DC, Boston, Albany and Syracuse.

The MTA has argued that countdown clocks at bus stops provide marginal benefit to riders at relatively high costs, and is focused on rolling out its BusTime program citywide by the end of next year. By that time, Lander would like a plan for bringing countdown clocks to the city’s 3,300 bus shelters. The route to achieving that goal is murky; Lander introduced the resolution to start the discussion.

Lander said that, ideally, revenue from advertising on countdown screens would fund the installation and maintenance of the clocks. If advertising could not cover all costs, he suggested they could be borne in part by Business Improvement Districts, council member discretionary funds or other local partners interested in bringing clocks to their areas. Lander added that bus countdown clocks were a popular idea during the last round of participatory budgeting in his district.

I have a simple challenge for Brad Lander: He feels it will cost up to $20 million to install countdown clocks that can use the BusTime API to pull data (although reliably converting distance to time is another issue). He believes it will cost around $5.28 million for maintenance each year. Instead of passing a symbolic resolution urging a solution, simply find a way to pay for the clocks. The $20 million outlay is hardly a huge expense, but it’s one the MTA doesn’t want to and cannot fund right now. Lander could.

It’s easy to talk about countdown clocks, but there’s been a concerted and logical effort to adopt a tracking system for a reasonable amount of money that doesn’t use clocks. Ideally, timers would be prevalent, but they are less reliable than GPS-based distance measurements. Now, though, a group of politicians want transit improvements at a concrete cost. Deliver the funding, and the clocks could become a reality.

Categories : Buses, MTA Technology

71 Responses to “For the MTA, it’s BusTime vs. Countdown Clocks”

  1. Christian says:

    We don’t need countdown clocks. What we do need is more Off-Board Fare Collections – Load times are paddle killers and unpredictable run separation can lead to 10+ minute load times, thus cripping x-town routes.

    • al says:

      At the very least, the MTA should deploy RFID fare cards to reduce the fare payment time to that of walking past the bus driver. Less time spent waiting for boarding passengers also result in quicker times between terminals. That can result in more passengers per operator/vehicle operation hr, greater bus frequency (shortening run time quickens return run in opposite direction, and so ad infinitum), and can improve farebox recovery ratios.

      They can also add off-board fare payment on commuter railroads to reduce on-board staffing, increase off peak train frequency (and ridership) and can improve farebox recovery ratios.

      • Matthias says:

        Both of these will be implemented with the next generation fare payment system.

        • Alon Levy says:

          No, they won’t. The MTA’s RFP explicitly says it assumes on-board fare collection except on the chosen SBS routes.

          • Matthias says:

            I was speaking to al’s comment regarding contactless fare payment on buses and in/out fare control for the railroads, both of which are planned.

            On-board fare collection would be just fine if there were all-door boarding and multiple validation points throughout the bus.

            • Alon Levy says:

              There’s no mention of all-door boarding on regular buses in the RFP. Nor is there a mention of turnstiles on commuter rail (which would be a stupid idea anyway).

              • SEAN says:

                No need for turnstyles with RFID cards as you “tap on & tap off” as you go. That is how you pay on CalTrain & it can work on LIRR, MNR & NJT trains just as easily. All that conductors need to check for proper fare is a hand held scanner.

              • Andrew says:

                The RFP calls for development of payment systems for both front-door-only and all-door boarding. Which of those systems to implement on which routes will be up to policymakers at the time of implementation.

                Proof-of-payment requires a large quantity of inspectors, and it would have been quite imprudent to assume in 2010 (or whenever the RFP was issued) that there would be enough inspectors on duty by 2015 (or whenever the full rollout happens) to handle all bus routes. On routes with few boardings at each stop, it might not be worth the investment in inspectors.

      • Josh says:

        The people who can’t seem to buy a Metrocard and need to pay their bus fare with nine quarters won’t be the quickest to adopt RFID fare cards either, I bet.

  2. John-2 says:

    For bus stops adjacent to places like candy shops, coffee shops or news stands that in tend to share customer traffic with buses, the MTA might consider some sort of joint agreement, in which the countdown clocks are actually in the windows of those stores, keeping them away from the elements/vandals, in exchange for the agency maintaining the clocks and partially subsidizing the wifi needed to make those devices work. Wouldn’t work at all stops, especially in residential areas, but you would be able to cover a significant number of stops in the system with that kind of partnership.

  3. Jeff says:

    I think the transportation blogosphere is being a bit pessimistic about modern technology’s ability to provide these time estimates given the bus’s position. Google’s algorithms for estimating time “in current traffic conditions” for their driving directions and turn-by-turn navigation are pretty advanced. They take into account real-time data collected from users of Android GPS devices heuristically blended with a whole host of other available data. It wouldn’t take a genius of a developer to tie this technology to BusTime and get us some pretty reasonable results.

    • How do you account for variable dwell times when the bus is picking up a larger-than-normal number of passengers at popular stops?

      • Tower18 says:

        You don’t have to. Obviously if the bus gets delayed, it stays on 5 minutes for a couple minutes, or whatever. But once the bus moves again, it’s 5 minutes away. Or it could pretty easily change to “delayed–approximately 5 minutes” if the bus has been stationary for more than 90 seconds, for instance.

      • Jeff says:

        This could certainly be accounted for (albeit imperfectly, as with any AI) in a learning algorithm. The algorithm could observe differentials in the bus’s travel time history in a given location and time of day/week compared to typical vehicle travel time in a given location/time. This would account for dwell times and other factors which cause a bus to travel differently than a typical vehicle, and be factored into the calculation of when the bus will arrive.

        Again, certainly not perfect, but the point is that people are going to use BusTime locations to come up with their own estimate for when the bus is going to get there (this is, after all, the whole point of bus time). So instead of individual users coming up with their own rough estimations, why not use the data and computation power available to come up with a better number?

        Does my faith in the possibilities of these algorithms necessarily justify countdown clocks at bus shelters? Not necessarily. That’s a much bigger question. As a user of the transit system, however, I can tell you that I would be much, much more likely to take the bus if countdown timers were available, albeit with imperfect numbers.

      • Alon Levy says:

        You fire everyone who thinks on-board fare collection is acceptable policy for a major city and try again.

        • Miles Bader says:

          Surely you meant “… for major destinations / transfer points.” Turning every teeny bus-stop in a city into a “bus station” would be absurdly expensive…

          [They use on-board collection everywhere here, and as far as I can see, it works fine, even at popular and busy stops (e.g. at stations). However most of the transit heavy-lifting is done by trains, so maybe the story would be different in a city with no reasonable rail system...]

          • Alon Levy says:

            In Hong Kong they also stick to on-board fare collection, reasoning that Octopus is fast enough. In Hong Kong buses are more important than in Tokyo, but their importance is declining relative to rail.

            But in Singapore they figured that smartcard validators are cheap enough they can put them at every bus stop and at every bus door for minimal cost.

            And then there’s the smartcard-free German-speaking world. When almost everyone has a season pass, the people who need to buy tickets can just board from the front and pay the driver and it’s a small deal.

    • Andrew says:

      I agree that it could be done, and while it wouldn’t be perfectly accurate, it would be close enough.

      The real reason only distances are shown? Most likely, Martin Kelly Jones.

  4. Someone says:

    Bustime is more reliable than countdown clocks. You never know when an accident can happen with countdown clocks but with Bustime you know exactly where the bus is. Even if there is traffic, you know whether the bus is close or not. With countdown clocks, you have to take into account that the speed of crosstown traffic may vary by street.

  5. Jonathan R. says:

    Great idea for those low-frequency buses like the B69, but those low-frequency buses also don’t have that many passengers. Lander’s proposal looks like a program whose target user is a rider who has no phone whatsoever yet is concerned about wasting time at a bus stop. How many people like this live in New York City? If there are 10,000 of them I would be surprised.

    • Jeff says:

      Smartphones and texting phones are not something that the primary demographic of bus riders – ie the elderly, readily have access to. So I would say your estimate is way off.

      • Jonathan R. says:

        The second condition I mentioned is “concerned about wasting time at a bus stop.” It’s been my observation that punctual people of any age generally carry phones.

        • Miles Bader says:

          I think “Jonathan R’s observation” is not a sufficient basis for making public policy.

          Anyway, even if there are only 10,000 people without phones—plus all the random other people that have their battery die, phone stolen (my friend in NYC just did!), forgot their phone, etc—saying “eh, screw you guys, buy an iphone already!” is the kind of position that’s almost guaranteed to get a politician in hot water.

          Countdown clocks are a nice extra, of course, not the transportation itself, so they could probably spin it by saying something like “we’re going to first roll out the cheap-n-easy service on smartphones and then add physical countdown clocks as best we can.”

      • Andrew says:

        The primary demographic of bus riders is the elderly?!

        Maybe on routes that closely parallel subway lines, but certainly not citywide.

        • Bolwerk says:

          Maybe saying the elderly is the primary demographic is an exaggeration, but I suspect the core demographic might be the elderly and a captive audience of people who have no other choice, and that audience might be largely socioeconomically poor. Combine that with the elderly and it certainly is a concern.

          OTOH, txt msgs r not goin away. Start now and the elderly will be using them in a decade or so. (I already txt my father a lot, and he’s near AARP age.)

  6. Patrick says:

    Even if they had the money, not every bus stop has a shelter, not even most high-traffic stops that will need the clocks.

  7. Someone says:

    What about building crosstown monorails where the crosstown buses are most congested? That uses off-board fare collection (mostly), uses its own tracks and is fully automated, which reduces travel times by up to 50%.

    • John says:

      Effectuating that plan is about as realistic as providing crosstown magic carpet rides.

    • Bolwerk says:

      Buses can use off-board fare collection just fine.

      Well, some kind of rail would be smart, especially if they could automate it. Though my “conspiracy theory” (for lack of a better term) is that the MTA’s love of buses largely boils down to the fact that they’re the most labor-intensive mode of all. Hell, it might explain why off-board fare collection is rare on buses too.

      • Alon Levy says:

        I don’t buy this. First, management dislikes the union, and loves to be able to frame anti-union measures, both justified and unjustified, as technological progress. And second, the union doesn’t even want bus drivers to handle fare collection, because of concerns for the safety of bus drivers who try to get unruly passengers to pay. If buses run faster because of off-board collection, the union has enough power to make sure service hours are maintained and frequency is increased, rather than that service hours are cut without loss of frequency.

        • Bolwerk says:

          First of all, I don’t think I buy management loves to frame anti-union measures for the sake of doing it. My impression is more that they just want the union to not bother them, which creates a strong incentive to capitulate to the union when they can afford to. The backdrop to that is an against an organization of type As, which surely doesn’t help.

          Second, I only said “might.” I wouldn’t rule out intra-union divisions (and/or intransigence) on this one. It’s very likely the official policy is the popular one, but the union leadership may well prefer to keep things from changing because, well, one change leads to another. I dunno. Again, I only said “might.”

  8. TERRANOVA47 says:

    If the system is anything like the one used by tfl, London Buses, in the UK, we will be watching a countdown of buses that approach, then disappear once due!

  9. SEAN says:

    In Portland Oregon, Tri-Met’s transit tracker has two formats. The original setup includes displays that tell riders how far the bus or train is from a given location via a stop ID number. You can call the automated info line as well. More recently, this information has become avaleable via text messaging.

    The way it works is via a transmiter that continuously relays the posission of every bus or train that is in revenue service. For a bus or train to be recognized, the opperator must insert an electronic card in to a slot on the dashboard that sets the route & destination. There was a video on youtube that explaned this system.

  10. Boris says:

    The point of making all the data public and providing an API is to allow anybody with a history of high-tech development better than MTA’s (e.g. everybody) to develop their own system. Any tech startup CEO with half a brain can develop an algorithm, rapidly prototype the system, and strike up a flexible ad-supported distribution scheme.

    Or, have the bus shelter company develop and install it – they are already in the advertising business and have a lot of control over their shelters.

    • Andrew says:


      And while I certainly wouldn’t object to having displays at bus stops, I’d rather know before I leave home when to expect the next bus, so I can leave just in time to catch it or walk to a different line if it’s going to come sooner.

  11. LLQBTT says:

    So no bus shelter = no countdown clock. Well that’s an awful lot of $ for a solution that covers only a small fraction of bus stops. BusTime has its flaws, but at least it will be universal.

    • Alon Levy says:

      Every bus stop should have shelter.

      • Phillip Roncoroni says:

        The problem is that there’s a bus stop every two blocks in Queens, which is unnecessary and slows things down. If bus stops were every ten blocks or so, then yes, there should be a bus shelter at every stop.

        • Phantom says:


          Spot on – same thing in Brooklyn. The bus stops are way too close to one another on many lines. It slows things up terribly, and ultimately it drives people from the buses.

        • Alon Levy says:

          I think it’s on Human Transit that commenters pointed out that if bus stops had shelter and seats, elderly people would be willing to walk longer knowing they’d be able to sit while waiting, and this would allow eliminating bus stops.

          • Phantom says:

            I dunno. Politicians listen to the .00000001% of people who complain about anything and everything, and its these no good lousy bastards who would complain about ” fewer stops “even the buses ran faster and if the remaining stops had heat in the winter and AC in the summer and water fountains with San Pelligrino all the year long.

          • Bolwerk says:

            Is there really research that frequent bus stops are that evil? Even at rush hour, NYC buses rarely appear make all stops.

            • Alon Levy says:

              Not that I know of, but in places where the usual design standards and tradition call for 400-500 meters between bus stops, transit ridership is higher than in North America.

              When I ride crosstown buses (a different animal from the north-south buses, which I don’t think I rode 10 times in 5 years of living in New York), they almost always make all stops.

              • Nathanael says:

                Moving sidewalks might be more effective than crosstown buses. (Channelling Hong Kong’s elevators here…)

              • Bolwerk says:

                Yeah, I just always chalked the manageable NYC bus issues up to boarding procedures and traffic. Distances between stops could very well be an issue too. Most of my present acquaintance with buses is with outer borough ones. There are often frequent stops, but they are rarely used by every bus unless there is a significant delay. Of course, you can make the case that even that’s still too much stopping.

                As for other places with more usual design standards, they usually don’t obsessively inflict buses on people the way North America does. :|

            • Phantom says:

              Research? Take the B63 ( Fifth Avenue ) bus in Brooklyn any day and see how painful it is with the constant stopping, which, combined with the lights, means that you’re not moving more than you’re moving.

              I can understand why many NYers, after a few experiences like that, simply never consider the bus as an option.

              • Patrick says:

                You know, I took that bus 2 weeks ago. ALL that stopping, freaking driver stoped at stops with no one at it. I think that route did slightly better running down Fourth Avenue

                • LLQBTT says:

                  it’s an OK line for a neighborhood or 2, but ride any farther and it becomes very tedious. Better off on the R for that even though it means more walking.

              • Bolwerk says:

                That’s not research. That’s an anecdote. The B63 should have more buses and light preemption. Busier buses of course deserve longer stop distances, and even to become LRT.

                But some buses should just be buses. I’m just not sure lower-traffic routes need to have their frequent stops eliminated. Maybe they should, and I’d buy it if the evidence pointed in that direction, but I just don’t know.

  12. Someone says:

    So GPS off = no Bustime. There may actually be more buses than is shown on the MTA’s Bustime website

  13. John S says:

    Curious point I seem to recall from the Problem Solvers event on BusTime – it seems that MTA’s hesitance to install devices is due to ADA requirements. Reading between the lines, it sounded like the idea was “If we install it, we’re bound by ADA to provide access to the blind, but if someone else wants to put up a screen using our data, have a blast.” I expect that if a BID (as suggested in a comment above) or a local vendor put a display in his/her window), this could be set up for not a lot of money in rather short order. (Inside, it would really just take a ~$300 PC plus a ~$500-$1000 monitor and a web browser locked on a BusTime web page. Outside, setting up a ‘robust’ computer, monitor, and Internet connection would add considerably to that cost – ~$8k, I expect.)

  14. Christopher says:

    A no talk of displays inside the bus to tell what street is coming up. Which seems to be a pretty basic part of systems before they have ever added bus tracking.

  15. Michael says:

    I commend the MTA for trying to use open source solutions rather than becoming dependent on expensive closed sole source alternatives.

    If the distance based bus time is open source, then why cant the MTA hire some programers and buy some computers to process the the time based travel data through the different stages of the day and week and try to develop a basic time prediction algorithms for each bus line?

    Ultimately all the systems out there just use historical data for travel times at different times of day and if really advanced some form of traffic data. Once the MTA collects the data, they could predict how long a bus will take to go 1 mile most of the time.

    As for displays at stops, maybe communities should take an initiative in installing and maintaining displays. if the data is on the internet, then anyone with a smart phone would be able to check arrival times, and if an advertiser wishes to sponsor a bus stop, they could provide advertising along with bus arrival times.

    Once the bus data is freely available, there are plenty of ways to share the information with riders. The key is processing and understanding the patterns in the bus travel data to develop predtion estimates that are roughly right most of the time.

    • Jacksonville says:

      They can hire Clever Devices back (M34/M34A SBS demo)… it was a little more effective then the open-source. Hell, NJTransit is using Clever Devices on its buses now.

      • The cost for Clever Devices and its closed nature (similar to MetroCards & Cubix) makes that idea both terrible and a non-starter. It’s possible to argue that no one should use Clever Devices.

    • Boris says:

      Like I mention above, *anyone* can hire some programmers and develop an algorithm. Perhaps the MTA can run an app competition for this. Historical data can be easily provided.

  16. Frank B says:

    Listen, even my 78 year old aunt in Marine Park has internet access to go online and look at bustime, ok? Before she goes outside to wait for the B2, she’ll click on the favorites bar, see when the bus is coming near her stop, and decide when to go out and wait.

    No hand-holding. BusTime is cheap, incredibly efficient, easy to use… even if you can’t figure it out on your phone, many seniors have internet access. BusTime is better than mere countdown clocks, because you have the information before you even leave your house; you can have that extra piece of toast; drink that extra cup of coffee; the bus is still 20 stops away.

    These countdown clocks will be huge targets for vandals, and largely redundant with BusTime. As much as I LOVE countdown clocks in the subways, I’d much prefer SubwayTime on my phone to tell me if the train is running on time before I leave work; maybe I have time to pick up a magazine on the way home.

    Two different systems; two different solutions. Forget countdown clocks for buses.


  1. [...] blue lights that help passengers identify Select Bus Service (SBS) buses are going away, another lost amenity for MTA transit [...]

  2. [...] ridership we’ve seen over the last few years. The debate, however, between BusTime’s location-based tracking and countdown clocks remains a hot [...]

Leave a Reply

XHTML: You can use these tags: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>