Greater Greater Washington. The Washington, DC area is great. But it could be greater.

Transit


Sometimes, less is more from trip planning tools

by Kurt Raschke   •   February 15, 2012 10:39 am

Matt Johnson argues that transit trip planning tools should show riders a wider range of options, illustrating how the schedules of connecting services (like bus and rail) mesh. That's often true, but for for a transit system with high-frequency routes, the best way to improve the usability of transit may be to show fewer times, not more.


Photo by Johan Rd on Flickr.

If a person is traveling between two points that are served by a high-frequency grid of routes, then what does it matter when they are leaving? When you provide a rider with a rigid itinerary—"here's how to get there if you leave at exactly 5:17 PM"—you give them the impression that if their departure time changes, then they have to re-plan their entire trip. With high-frequency routes, that simply isn't the case.

If a rider can take the trip entirely using high-frequency routes, doesn't it seem so much more liberating to tell the rider to "show up any time and arrive within 45 minutes"?

Simplifying directions like this helps riders internalize the route network, and encourages spontaneity. Instead of having the sense that every transit trip starts with a visit to Google Transit, riders gain the sense that they can travel whenever they want.

Jarrett Walker emphasized the value of grids, and of high-frequency transit services, during his talk last week. "Frequency is freedom," he says. A regular grid of frequent services makes it easier to get around without having to consult an online trip planner before every trip, though many riders still rely on Google Transit and local trip planners to figure out how to get around.

In fact, the worst thing a trip planner can do is recommend that a rider take an infrequent, irregular service just because it happens to be there when the rider is starting their trip. A great example of this is the Route 305 bus in Los Angeles; as Jarrett Walker explains, it's a low-frequency service which runs through a high-frequency grid:

That means that the 305 is the fastest path between two points on the line only if it happens to be coming soon. If you just miss one, there's another way to get there faster, via the much more frequent lines that flow north-south and east-west across this entire area.
Why should a trip planner ever recommend that a rider take a bus like the Route 305? Doesn't it make more sense to show them to how to use the high-frequency grid to their advantage?

Our hapless, misdirected rider will doggedly wait for that infrequent route to come along, because it's what their itinerary lists. But if they'd received an itinerary which used the high-frequency grid, they'd be on their way a lot sooner.

Of course, even in cities with the most comprehensive high-frequency grids, some trips require going outside the grid. Then, there may be no choice but to ask the prospective rider when they're travelling. But even in those cases, the trip planner's itinerary should still include information on the frequency of the services being used.

Simply put, out in the real world, things happen. A rider might get to the bus stop or train station 10 minutes or even 30 minutes after they'd intended, so doesn't it make sense to tell them up front how long they'll have to wait if they miss the planned trip?

Cross-posted at Raschke on Transport.

25 comments
Tags: buses, Jarrett Walker, transit, trip planning
Related posts:
Kurt Raschke is an information technology professional and transit enthusiast interested in how technology can improve the usability of transit systems. A car-free resident of Silver Spring, he is a frequent user of Metrorail and Metrobus. He also blogs at Raschke on Transport. All views expressed here are his alone. 

Comments

Add a comment »

So are we deeming DC to be a high-frequency grid? I have a hard time buying into that idea. A subway system that sometimes has 20-minute headways is not high frequency. Given the sporadic schedule of local transit(peak times on metro, weekday only bus lines, etc), I believe that Matt Johnson is right. Real time information is most helpful. In fantasy world, there would be a bus or train every two minutes. DC is very far removed from fantasy world. Most of us are happy if the train shows up at all and doesn't break down on our trip.

by MJ on Feb 15, 2012 10:50 am  (link)

A subway system that sometimes has 20-minute headways is not high frequency.

Which is why we should close Metro at 10, and run buses every 5 minutes along the same route afterward.

There *are* high-frequency corridors in the city, but only a few of them. 14th St NW is the most notable one. During most hours of the day, you rarely ever have to wait more than 5 minutes for a "regular" bus or 15 for an "express" (14th St Circulator) bus.

For this reason, I usually opt to take the bus at night instead of Metro, as long as I'm traveling along one of the frequent corridors. Otherwise, I'll use NextBus to time my departure for a bus route that doesn't require a transfer.

Honestly, I wouldn't mind if Metro scaled back its hours during the "reconstruction" as long as it pledged to run a good substitute bus service. It'd probably actually reduce travel times for most folks. London does this all the time.

Metro is failing as a rapid transit system if it is unable to offer more than 20-minute headways, and needs to provide other options for frequent service in town.

by andrew on Feb 15, 2012 12:36 pm  (link)

Most of us are happy if the train shows up at all and doesn't break down on our trip.

Cut the hyperbole. This isn't a real concern that people actually have.

by andrew on Feb 15, 2012 12:39 pm  (link)

@MJ

I'm afraid you're a bit off the mark. I didn't make any claims about the mass transit available in DC, nor did I recommend concrete changes to any specific trip planners.

Instead, I presented a conceptual framework for considering trip planning in high-frequency grids.

As to whether or not such a grid exists in the area, I would argue that it's fairly obvious that there isn't one. This is something that Michael Perkins has been pushing for, along with pass reform, and I wish him luck in that endeavor.

@andrew

Agreed. I've long considered it strange that Metrobus doesn't have a system of "night buses" as is common in other cities which shut down their main transit routes (whether bus or rail) at night.

by Kurt Raschke on Feb 15, 2012 12:57 pm  (link)

I'd rather take the 305 than muck around with transfers. I just need to know when to be there.

by movement on Feb 15, 2012 12:59 pm  (link)

I think we need to re-frame the trip planning question.

What I'd like is someting interactive; ie a smartphone app much like a GPS app. Smart enough to know if I should walk a few blocks to catch a better bus, or just wait. I do this myself comapring the 38B, 3y, and Circulators. You've got to intergrate traffic conditions in there. When there is a jam up to the bridge, the 3y is a nightmare.

Hidden benefit of nextbus -- getting more middle class people on the bus. I suspect that would be even larger with a trip planning app.

by charlie on Feb 15, 2012 1:58 pm  (link)

@charlie

Why use GPS as the example? GPS is making us dumber. It hinders our ability to navigate.

Instead of relying on an app to navigate a complex system, we ought to endeavor to make the system simpler, which forgoes the need for an app in the first place.

by Alex B. on Feb 15, 2012 2:09 pm  (link)

I wouldn't mind if Metro scaled back its hours during the "reconstruction" as long as it pledged to run a good substitute bus service.

Hahahahahahaha. Just how reliable would you expect such a "pledge" to be?

by Marian Berry on Feb 15, 2012 2:14 pm  (link)

@AlexB, actually, I agree -- GPS based navigation (not GPS) does make people dumber. I've seen it myself.

But it also enables and opens up mobility for more people. Micro vs. Macro.

From my example, no about of scheduling would help determine which bus will get your from the west end to arlington faster. It is variable -- traffic, delays, crowding -- all play a role.

by charlie on Feb 15, 2012 2:22 pm  (link)

I thought GPS is much like the internet, it gives us access to information we might not otherwise have or at least w/limited success.

If the current Metro apps operate on a GPS system (doesn't it?) then I absolutely love the integration as it makes my commuting better and I'm likely to do it more often outside of the 9-5.

by HogWash on Feb 15, 2012 3:02 pm  (link)

Charlie,

Yes, GPS navigation. Advanced trip planners are more like GPS navigation.

Just plain GPS is nothing more than a fancy map.

by Alex B. on Feb 15, 2012 3:09 pm  (link)

@charlie

In places where the transit agency has provided google with real-time data (rather than just schedule data) the trip planner can sort of do what you're saying.

But it still can't update in real-time how long it's taking the bus to get to your destination.

by MLD on Feb 15, 2012 3:09 pm  (link)

@AlexB, maybe we're talking at cross purposes, but GPS is a technology. That's it. GPS enabled maps? GPS based navigation? All slighly different.

@MLD; nextBus can do the same. Absent ghost buses, you can actually sense distrubtance in traffic. Well, not on the 3y, because it runs every 40 minutes or so, and a traffic jam can build up then.

Another example is people who check the next train to see if it makes sense to hop of the red line, then transit to the orange line at metro center -- rather than walking 5 blocks to metro center directly. Real time feedback.

by charlie on Feb 15, 2012 4:03 pm  (link)

Charlie,

maybe we're talking at cross purposes, but GPS is a technology. That's it. GPS enabled maps? GPS based navigation? All slighly different.

Right. What I meant was GPS navigation (which I would contend is what most people mean when they utter 'GPS.')

Another example is people who check the next train to see if it makes sense to hop of the red line, then transit to the orange line at metro center -- rather than walking 5 blocks to metro center directly. Real time feedback.

I do this, too.

The point is that if the services are running frequently enough, it shouldn't matter.

Instead of building an app so that people can do this kind of wayfinding on their own, structure the service so that kind of wayfinding isn't necessary in the first place.

by Alex B. on Feb 15, 2012 4:11 pm  (link)

Another example is people who check the next train to see if it makes sense to hop of the red line,

Or hurry (or not) to get the next train as you're walking towards the station.

It really kicks a**

by HogWash on Feb 15, 2012 4:33 pm  (link)

One thing Metro's trip planner doesn't do is give us the next itinerary with an arrival after the specified arrival time. Often if I'm going to the airport I want to get there an hour before the flight (so I put in, say 6:20pm as arrival). It will give me a list of trips getting there before 6:20. Great. But if there is a train that gets me there at 6:22, I'd like to know that - also if the next one only gets me there at 7pm.

This is information that the current system makes hard to get. I'd like the planner to give me the next connection AFTER the time I'd like to get there, as well as the ones before.

by egk on Feb 15, 2012 5:21 pm  (link)

Actually, the 305 is an interesting example, and this discussion is a bit more relevant to me than I previously thought.

I spent a few days last week in San Francisco. I came away with a few thoughts about the transit there:

  • Their map is almost (but not quite) as bad as the DC Bus map. Worse still, it's very difficult to distinguish different modes of transportation on the map. BART's not even on there, but the slow $6 cable car is. I think I used it twice.
  • They've got much more frequent service on most bus lines. Apart from the lack of bikesharing, it's much easier to get around within the city than it is in DC. I couldn't actually identify many priority corridors because of the terrible map, but could walk up to most bus shelters, and see an electronic sign indicating a bus or tram arriving in less than 5 minutes. This is huge, and DC needs it as a reassurance regardless of whether or not we're also going to have frequent service.
  • They advertise their awesome mobile site (link doesn't work in a desktop browser), which helps pedestrians select the best spot to board the bus. It's location-aware and refreshes automatically, giving it app-like functionality. Ironically, the same site works in DC, because it pulls its data from NextBus. Why don't we promote this? The desktop version of NextBus is already buried on WMATA's site, the dial-in service is terrible, and the SMS service is completely unadvertised.
  • Speaking of smartphones, the Google Transit Android app is incredible. Incredible enough to begin to undermine Kurt's point (when it comes to daytime service). I think we could argue that, as long as we've got priority corridors with frequent service, it's fine to continue running obscure routes across town as long as the ridership's there to support them. For one of my trips, Google Transit suggested taking an obscure, once-hourly bus across town that would eliminate the need to transfer, cut about 20 minutes off of my travel time, and literally take me door to door. I mean.... why not? Unlike NextMuni/NextBus, it also suggests routes from all of SF's transportation agencies. Despite its bad rap, BART's still useful for certain in-town trips. The app even makes your phone vibrate as you're pulling up to your stop; how cool is that?
  • There need to be more places to add fare to fare/smartcards, or buses need to accept more than $1 bills and quarters. Every vending machine on the planet can make change, and many can even take cards now. Buses need to catch up. If SF didn't offer paper transfers, instead requiring a difficult-to-acquire smart card like DC does, I'd never have used the bus.
  • In spite of these sometimes-handy cross-town routes, DC certainly has some completely bizarre routes that could be eliminated or merged. How do you explain the X3? It runs 7 times a day Westbound, and 5 times Eastbound (a riddle in itself) along about 90% of the same route as the 90/92/93. While we're on that topic, the 90/92/93 are similar enough to be potentially treated as the same route.

by andrew on Feb 15, 2012 5:51 pm  (link)

Oh. Sorry. I should clarify that Kurt's point is undermined because it makes far more sense to use a trip planner on a mobile device.

We're spending too much time worrying about desktop-based trip planners, which are far less useful to the majority of transit-users. A simple map, priority corridors, and a smartphone app for people seeking efficiency are all far more valuable than a good online trip planner (which Google already does better).

by andrew on Feb 15, 2012 5:54 pm  (link)

@andrew

So are you going to buy every rider, now and in the future, an Android device?

(and, not to go off on a tangent, but that's assuming they'll use it; I am a loyal Symbian (make that Nokia Belle) user, and will be until the bitter end)

In all seriousness, it's not an either-or. In fact, from a perspective of equity (that is, ensuring that transit is accessible to as many riders as possible), presenting passenger information in a variety of formats is valuable. Some riders may benefit from a real-time trip planning app. Others may want to plan their trips in advance, and prefer a format like Matt Johnson's recent prototype. Still others may have irregular schedules, and prefer directions that emphasize routes over times, giving them a sequence of routes that will get them where they're going no matter when they leave (which, of course, only works in a city with a high-frequency grid). All of the above can be accommodated by a sufficiently flexible trip planner (particularly one with an open API that encourages developers to innovate).

As to your point concerning the "obscure, once-hourly bus across town", I think you're confusing the situation somewhat. NextBus is not a trip planner. As I wrote in the post, "some trips require going outside the grid". This includes trips where the high-frequency grid provides the necessary service but is for whatever reason substantially less convenient. If the direct route really does save a rider that much time, then show it as the first option—but, again, "the trip planner's itinerary should still include information on the frequency of the services being used". It's fortunate that you caught your bus on time, and that if you hadn't, you'd have been able to re-plan your trip on a mobile device. But not everyone will be so lucky. So, for those riders, what's the harm in having the trip planner's itinerary include a warning that the next bus doesn't come for an hour? Then it can offer the user alternatives, where they exist—and if the only option is a low-frequency bus, then at least the user is forewarned.

by Kurt Raschke on Feb 15, 2012 8:14 pm  (link)

@andrew

A few points:
1. Thanks for the NextBus site tip - I've never used the mobile site before and didn't know that functionality was there.

2. The bus doesn't make change because it would slow down boarding. Can you imagine how long it would take to get people on the bus if, in addition to people using cash slowing it down, those people had to take more time to gather their change? Boarding is the biggest thing slowing the bus down. You have to make it as fast as possible.

3. As for smart cards being "difficult to acquire," you can go directly onto the Muni website and order a clipper card with your credit card. They mail it directly to your house. You don't have to go anywhere. And their card is even more convenient than the WMATA one because it has passes AND can autoload from your credit card when it gets low. I can't believe you think a paper transfer is more convenient, especially since your target audience owns a smartphone.

by MLD on Feb 16, 2012 8:07 am  (link)

@Andrews; that mobile site is nice! I am so tired of trying to pinch down on the nextbus sites to grab a line. However, you've got the same problem with the main WMATA site - unless you bookmark the list of buses is too long. I can't get the location aware to work on the iphone, though.

Equity is a nice goal. But in reality, what buses need are more middle class users and mobile apps are a good way to bring that.

by charlie on Feb 16, 2012 9:07 am  (link)

I think this would be a more compelling counterpoint to Matt Johnson's article if it could make its argument for the DC area, as his article did. I don't think we have enough higher-frequency routes for the "show up at 9 and be to your destination by 9:45" way of thinking to really work.

As theory, yes, good article. Applied to DC, where our headways on even metro service through the core can be 3 minutes, 6 minutes, 12 minutes, 20 minutes, or sometimes pushing 30 on the weekend, a little harder to master.

by worthing on Feb 16, 2012 9:25 am  (link)

@charlie: If you grant the site permission to access your location, you shouldn't need to select the bus from the long list. About 50% of its beauty lies in that fact alone.

And, yes. I totally understand Kurt's point that not all of us have smartphones at our disposal. However, I wouldn't underestimate the proliferation of technology. We were saying the same thing about mobile phones just a few years ago, and even the low-cost carriers are jumping on the data bandwagon.

As you said, having a variety of options at our disposal is by far the best (albeit most expensive) strategy. Priority corridors are valuable for impromptu trips, and lower the "barrier to entry" for transit for more folks than an IT solution ever will.

A trip planner and bus map that properly highlight these corridors as "reliable safe bets for a quick bus" definitely improve the effectiveness of said priority corridors.

...however, IT solutions are pretty nifty too. Google Maps does a better job with transit directions than WMATA's planner does, and I would hope that open data allows the worldwide community to continually improve these applications and algorithms for all agencies via the Open-Source community.

The Google transit app made it *dead* simple to say "I want to get to [X] from here," and be presented with the best way to do so now, 5 minutes from now, and 15 minutes from now. It's not a catchall solution, but I was surprised by just how much more I relied on transit in an unfamiliar city, thanks to this easily-accessible and reliable information.

If we're concerned that the audience of smartphone users is too small, we could build a trip planner with a voice-recognition dial-in system or SMS gateway. (The NextBus dial-in line was terrible the last time I used it about a year ago)

But, again. Transit users shouldn't need to rely on an app for a quick impromptu trip. That's where priority corridors and real-time arrival displays come in.

We also need to keep traffic flowing on those priority corridors. The Circulator usually manages to do a great job with this, while WMATA still faces persistent issues with bunching. Although traffic and bad signal timing contribute to this, WMATA still has some poorly-timed departures and tight turnarounds on its schedule that could easily be corrected. However, Metrobus has never been a priority for the agency.

by andrew on Feb 16, 2012 10:20 am  (link)

@andrew

I can't find a "Google Transit" android app - are you talking about the "Get Directions" function of the "Maps" app? I find that to involve a lot more menu navigation than I would want from something I relied on for quick transit information.

by Lucre on Feb 16, 2012 11:27 am  (link)

@Andrew; I'm not getting on option on the mobile site to allow location services; perhaps it is android only feature.

But just the navigation is 10x better than the older wmata site.

@Kurt; I was an old Symbian hand as well but welcome to th new world.

by charlie on Feb 16, 2012 11:35 am  (link)

Add a Comment

Name: (will be displayed on the comments page)

Email: (must be your real address, but will be kept private)

URL: (optional, will be displayed)

Your comment:

By submitting a comment, you agree to abide by our comment policy.

Notify me of followup comments via email. (You can also subscribe without commenting.)

or see below to post

To post your comment, please enter the two words in the box below to prevent spam:

Save my name and email address on this computer so I don't have to enter it next time