Frequent Networks

“We can’t have density there because there’s not enough transit.”

Ever heard this line?  A debate in Google's home town, Mountain View south of San Francisco, has turned up this response to an obvious idea of building more housing close to the city's business-park district, so that fewer people have to drive long distances to get there.  No, some council candidates say, because there's not enough transit there. 

Well, there's not enough transit there because there aren't enough people there, yet.  Transit is easy to add in response to seriously transit-oriented development, but as long as you have a development pattern that is too low-density or single-use for transit, you've locked in lousy transit service as an outcome.  

So whenever someone gives you this line as a reason to oppose a transit-friendly development, ask: "Well, what would it cost to provide good transit, and who should pay for that?"

Often, as in Mountain View, extremely frequent transit into the nearby transit hub can achieve plenty, and is not that expensive, because of the very short distances involved.

There are other situations where there's not enough transit because transit just isn't viable at any reasonable price, for an obvious geographic reason like remoteness from transit hubs or destinations.   

But it's worth asking.  

Frequent networks: escaping the “food desert”

For their piece on food deserts this week, National Geographic led off with this map of Houston.  It shows where large numbers of people who lack cars are located more than 1/2 mile from a grocery store.  

Houston food desert

"Public transportation may not fill the gap," the article says, but sometimes it can.  The article doesn't mention it, but Houston METRO's proposed System Reimagining will actually liberate many people, but not all, from the "food desert" problem.  

One thing is for sure: When we're talking about errand trips like groceries, most customers don't have a lot of time.  If there's a line to check out of the grocery store, and you miss an hourly bus, you'd better not have bought anything that needs refrigeration, and certainly nothing frozen.  So as always, frequency is freedom.  

So if you need frequency, Houston's transit system today is basically willing to take you downtown, which is not likely to be the path to your nearest grocery story.  Here's the frequent network today. 


It's radial, good for going downtown but not for many other purposes.  If you're lucky it will take you to a grocery store, but more likely it misses the ones nearest you.  And of course you may not be on this network at all.

Here's the reimagined frequent network:


Compare this to the food desert map (to see the detailed Reimagined network maps, see here)  Many of the areas of concern, especially those in the southeast and southwest, get a much richer network in a grid pattern.  The grid pattern means ready access to many commercial centers in your part of town, not just to the major destinations of the region.  And of course, a much larger share of the "food desert" areas are on this network, so that they can make shopping trips that take an hour instead of all afternoon.

Another crucial thing about the grid is that by running in all directions, it cuts across socioeconomic divides.  Low-income people can get out of their enclaves to reach both jobs and commercial services in more prosperous areas nearby.   (This frightens some people on the upper side of these divides, but it's one of the basic ways that good transit that's broadly useful creates paths out of poverty.)

(Remember: Most low-income people are busy!  They have to be frugal with time as well as money!)

Not all the "food deserts" can be healed with transit.  Some of the highlighted areas on the food desert map are in the northwest and northeast, outside of loop 610.  Low-income housing in this area takes many forms, but a lot of it is semi-rural, and the built environment is often very hostile to both pedestrians and efficient transit routing.  Some of these areas also have declining population.  Our plan does try to offer some options in these areas, but transit is not the primary solution to the food desert problem there.

But over a very diverse range of Houston, the way to get low-income people to decent healthy food is the way you achieve so many other benefits: environmental, economic, and social: an abundant frequent transit system, in a grid pattern, that reaches across all the parts of the city that are dense enough to support it.  

columbus: a new transit network plan

Columbus, Ohio's metro transit agency, COTA, has now released a new network plan for public comment.  As in the recently unveiled similar plan for Houston, I led the network design task on this project as part of a consulting team led by IBI Associates.  

Again, the core idea is to expand the Frequent Network — the network of services that run every 15 minutes or better all day — so that more people have service that is highly useful.  Here's the existing Columbus area frequent network :


And here is the Draft Proposed Frequent Network:

 In Houston, we achieved similar expansion solely by reallocating existing resources.  In Columbus, there was a small budget for expanded service, but still, 90% of what is achieved here is the result of reallocation: removing overlapping routes and deviations, removing duplication, and in some cases removing service that very small numbers of people were using.  

Details of the plan are on the COTA website, here.   The total proposed network is here.  Note that color denotes all-day frequency: red is 15, blue is 30, green is 60.  The plan does many other good things, including a major expansion of weekend service.  


 You can upload the existing network, for comparison, here:   Download Existing System Frequency Map

Again, if you're in the Columbus area, please comment to COTA using this special email:   [email protected].  At this stage there is no decision about whether to implement a plan such as this one.  Any final plan will be revised based on public comment that comes in over the next couple of weeks.  That means that if you like the plan, it's important to comment to that effect, as well.  


portland: the high-frequency grid is back!

FrequentserviceThe reputation of Portland as a transit city in the last two decades may have been about its light rail, streetcar, or aerial tram, but to the extent that the city achieved significant ridership and made transit a welcomed part of urban life, it's done so with a grid of frequent transit services.  

This grid consisted of both rail and bus services, but it's the frequent grid pattern, not the technology, that made it easy to get around the city, in ways you could measure with trip planning software.  For example, the amount of the city you can get to quickly from the inner-city Hollywood district arises mostly from frequent grid connections , as from the light rail line there. 


[Graphic by Conveyal showing travel times on transit+walking from Hollywood light rail station.  Blue is 15 min, green is 30, pink is 45.]

Back in 2012, this post celebrating the 30th birthday of the high-frequency grid ended with a serious caveat:

Purists might argue that the grid never made it to its 30th birthday, but rather perished at  age 27 in 2009.  That was the year that TriMet cut all-day frequencies below the 15-minute threshhold that is widely accepted as the definition of "frequent enough that you can use it spontaneously, without building your life around the timetable."  Since the grid relies on easy connections to achieve its goal of easy anywhere-to-anywhere access, the 2009 cuts began to undermine the whole idea of the grid. TriMet avoided doing this in its first round of cutting after the crash, but felt it had no alternative in the second 2009 round.  

Will the grid ever be restored to its necessary frequency?  Will it ever be expanded and enriched (as regional land use planning generally assumes it must be) with even better frequencies?  Not everyone in Portland thinks this is a priority, so you might want to express your view.  

Earlier this month, TriMet answered that question by restoring 15 minute service to 10 critical bus routes, in addition to the two routes which maintained that standard through the recession and recovery. During the period of this service cut, as good connections enabled by the frequent grid became more arduous and wait times increased, the overall utility of the system across its strongest market was diminished. The impact on ridership was clear: between fall 2012 and fall 2013 alone, weekday bus ridership declined 3.6 percent, 2.6 on the MAX light rail system, continuing the negative trend starting from 2009.  

The return of the grid is good news for riders, and no doubt the agency hopes to reverse the troubling ridership trend and create some good publicity in the processs. High frequency transit service is a key characteristic of many of Portland's most attractive neighborhoods, and must be seen as a permanent element of these places if they are to continue to grow in a manner that enables people to make real choices about their travel options. 

More importantly, the return of the grid is good for anyone who wants Portland to be a denser, more walkable, more sustainable city.  Portland policy allows lower minimum parking requirements for dense housing located along the frequent network, and those low requirements help make such development viable as an alternative to sprawl.  Last year, though, there was an eruption of controversy around these requirements in one inner neighborhood, and one of the legitimate objections was that the full span of frequency (7 days including evenings, as you need for a liberated life on transit) had been cut in 2009.  

The fall and rise of Portland's frequent grid shows the perils to actual freedom of access that arise when buses aren't respected for their essential role, and when nobody steps up in an economic crisis to save a crucial building block of the city's redevelopment policy.  Portland's 2009 frequency cuts drove away a lot of transit riders.  Please spread the word that they are welcome back.

the evolution of logic in privately planned transit

Step out into most developing world cities, and you’ll see something like this:


Lots of vans sitting around, looking like maybe they’re about to go somewhere useful.  Vague cardboard signs in the windows suggest they may or may not be public transit of some kind.

They’re called matatus in Kenya, colectivos in Latin America.  Over much of the world these informal, private, for-profit vans, run at low cost for low fares in areas of high demand, forming the basic public transit for a city.  Generally they run along a particular route out of a hub like the one above, but sometimes it’s possible to vary the route depending on what you can negotiate with the driver.  You can count on them to hit key locations but not necessarily the exact path they’ll take.  You also can’t be sure of when they’ll go.  Sometimes they wait until they’re full before leaving.

Today in Atlantic Cities Emily Badger tells the story of the Digital Matatus project, an attempt to map and describe the spontaneously evolved patterns that these semi-fixed-route buses operate.  Although nobody planned this network, it’s more orderly than you’d guess.  Download the sharp, complete map here.

Matatu map slice

What do I notice?  Practically everything goes downtown!

Matatus have organized themselves into routes because that’s to their benefit; they train customers where to wait for them along reasonable paths so that they aren’t driving around looking for customers individually.  The idea of the route — and of an efficient, non-duplicative spacing between routes — arises spontaneously from the economics of the product.

But they almost all converge on downtown, creating huge jams there.  Nairobi is clearly big enough to  have large flows of people crosstown to many non-downtown destinations, suggesting that a more efficient and liberating network would have more grid elements.  This is a common thing that goes wrong in privately evolved systems.  Every matatu wants to go downtown because it’s the biggest market, and a mutatu driver doesn’t have to be coordinated with anyone else to fill a bus going to and from there.  This geometry problem bedevils privately routed and scheduled operations everywhere.

Crosstown service, by contrast, requires frequency on a single path connecting several major dots, and it has to leave from organized non-downtown hubs where many other services connect to it.   That requires more organization, so it’s less likely to arise spontaneously out of private operators optimizing for themselves.

So you get a single market overserved and other markets underserved.  This is very much like the way a narrowly-focused transit agency will throw too much service at a single market rather than building a network useful for many markets.  It takes more planning and management to create a network, and this usually requires a government willing to impose order.

This same problem was observable after the wholesale privatization of buses in Britain.  Suddenly there was lots of duplication of bus service into the biggest downtowns as everyone chased the easiest prize, but service disappeared from crosstown markets that could have done well, but that required a network of organized connections to succeed.  That network is what privately motivated transit has trouble delivering, because it usually requires cooperating with people who are perceived as competitors.

Now and then, these systems get reorganized by government into more logical routes that spread the network across the city for easier everywhere-everywhere travel, as happened in Santiago in 2007.   The transition is hell, but when you’re finished, you have a network that’s much easier to use to go all over the city, and a much smaller knot of buses downtown.

The moral?  Disorganized transit systems “planned” by the actions of many private actors do naturally evolve certain forms of efficiency, but they do not naturally evolve into the most efficient and productive network for the whole city.  That final push into coherence requires network design!

new york:a frequent network map

Tumblr_inline_mq8zieIEsm1qz4rgpJust found this map of all 10-minute frequency or better services in New York City, by this not readily identifiable character on Tumblr.  This looks like quite a struggle to make clear given the complex nomenclature that NYCTA uses.  

The whole thing is here.  An NYCTA contact tells me it’s still current except for a change around LaGuardia airport.

When I’m learning a new city — as I do 10 or 20 times a year as a consultant — this is what I need!


should transit maps be geographical or abstract?

In some agencies, it goes without saying that transit maps should be geographically accurate.  Many agencies follow San Francisco Muni in superimposing transit lines on a detailed map of the city:

Sf frag

But research out of MIT suggests that we really need to see network structure, and that requires a degree of abstraction:

By putting alternate versions of the New York and Boston subway maps through the computer model, the researchers showed that abstract versions of the maps (as opposed to geographically accurate versions) were more likely to be easily understood in a single, passing glance. 

Here's their example:

Geographical accuracy obscures network structure.  Purely geographic maps show where service is but not how it works.  

This is why a number of best practice agencies publish both kinds of maps, sometimes even presenting them side by side.  The geographic map helps you locate yourself and points of interest in the city, but you need the structure map to understand how the system works.

All this is even more urgently true for bus network maps, where complexity can be crushing to the user.  When we streamline maps to highlight key distinctions of usefulness such as frequency, we often have to compromise on geographic detail.  Obviously the best maps fuse elements of the two, but you can always find the tradeoff in action.  The new Washington DC transit maps, for example, highlight frequency (and show all operators' services together) but there's a limit to the number of points of interst you can highlight when keeping the structure clear:  

Dc slice


frequent network maps: the challenge of one-way pairs

One-way splits — where the two directions of travel are on different streets — are often the scourge of transit: on the map, for example, they appear to cover more area than two way service, but actually serve less.  And they certainly make transit maps confusing:


Still, they're frequently mandated by one-way traffic couplets.  Those, in turn, are usually mandated by the goal of flushing traffic through a city, though there are cases, notably Portland, where one-way couplets are perfect for creating an intimate and walkable downtown.

Transit agencies may not be able to avoid one-way couplets, but they can control how they describe them and think about them.  WMATA (and its map designer, CHK America, have made a major step in their new network map.  We covered the development of this mpa previously. Here's what it looks like today:

Screen Shot 2013-09-30 at 10.22.19

Red represents the most frequent bus lines, blue the infrequent ones, and black the DC Metro system. Other colors are used to depict services of other transit agencies. This basic, clear symbolization quickly communicates the relative importance and usefulness of each type of service.

Marc Szarkowski, who contributed his own frequent network map of Baltimore to the blog last week, asks: how do we show a single route that runs as a two-way couplet on separate streets, without introducing too much clutter or confusion? 

Marc writes:

I think they can work if presented effectively, but overall I often find them confusing, especially if I'm taking an unfamiliar route to an unfamiliar area (all the more you have to remember, particularly if you take multiple such routes). For example, whenever I ride a bus to an unfamiliar area, I tend to assume that the stop I get off at in one direction is just as good for boarding in the other direction. It's frustrating to return to the stop just to discover that you have to walk a block over (or sometimes more: see the 10 in West Baltimore!) to catch the same route in the other direction. 

Screen Shot 2013-09-30 at 10.29.07

WMATA's map uses one line to show both branches of route, and labels either side with the couplet streets. This effectively reduces the amount of clutter on the map, but also excludes which direction the bus travels on each street. This information is less crucial in the case of the B2 shown left, where no other routes travel on the same pair of streets, but where multiple routes use the same streets, in different patters towards different destinations, combining paired one-way streets can become very confusing. Marc's map does not employ this method of simplification for the same reason: 

In Baltimore's case multiple overlapping routes were sometimes offset across a series of three or more one-way streets; i.e. Route 1 up on street A and down on B, Route 2 down on B and up on C, Route 3 up on C and down on D, and so on

Ultimately, desiging this type of a map is about balancing information density and comprehensibility. The user needs to know that line B2 runs on both 14th and 15th streets, but for a map at citywide scale, it may be more important to communicate that B2 is a frequent line serving a long, straight corridor on the eastern side of the city. The map already distorts direction and the exact shape of the streets in favor of a simpler visual effect. WMATA's map uses the same approach to one-way street pairs, downplaying accuracy in favor of ease of use.

frequent network maps: baltimore

Another unofficial frequent network map, this time via Human Transit reader and Envision Baltimore contributer Marc Szarkowski: 


Baltimore Frequent Network

You can see the full size version here. This map uses color to differentiate between linear and loop routes, and line weight to denote frequency. Compare this to MTA's current system map, showing the same part of the city:

Screen Shot 2013-09-26 at 11.45.04

Existing Baltimore MTA map

MTA's map uses a variety of colors to depict individual routes, but without assigning them significant qualities (though green and blue appear to be reserved for different kinds of commuter expresses). The seemingly arbitrary assignment of colors to local routes creates a cluttered, confusing visual effect, and obscures the quality of service provided by each route.

On the other hand, Marc's map distinguishes which routes on which roads provide which level of service, using a simple 5-color scheme differentiating linear and circulator routes, rail, rapid bus, and ferries.  Frequent service is clear as a dark wide line, with its color indicating technology.  This visualization is very information-rich,  offers a clear improvement to the MTA map in its utility as a description of the service available to MTA riders.The image below is a snapshot of the legend from this map:

Screen Shot 2013-09-26 at 12.12.27

Marc writes:

The radial nature of the network is also why I decided to group services into a limited color palette, like the WMATA map. Originally I intended to assign each route its own color, as in the Leeds or Portland maps, but so many routes crowded together in so many areas as they headed downtown that it was difficult to fit the whole "spectrum" on certain streets/corridors. (I think this is why even the MTA regional map, which does use a wider color range, still has to resort to using a single line for all buses in central Balto.) If the network was more emphatically organized on a grid, as you advocate, assigning individual route colors would be a lot easier since the map wouldn't have to display as many redundancies.

Screen Shot 2013-09-26 at 12.02.01 In the image at left, for example, the visual prominence of the blue lines for Route 5 clearly communicate its higher level of service, compared to Route 91 several blocks north. The directional arrows attached to the route labels are also a nice, unobtrusive touch; transit maps can often become unecessarily cluttered with these symbols, particularly in systems with many looping routes.  Marc's map does a nice job of providing this necessary information at a relatively low level in the visual hierarchy. 

Intelligently designed maps like this one show opportunities for connection, and the relative importance and usefulness of the system's transit routes These maps work by exposing the degree of freedom of mobility available to a transit rider.

real-time transit mapping: watch your system in motion

Many transit agencies subscribe to the idea of "Open Data", publishing bundles of information on the system. These feeds contain data on stops, routes, arrival times, fare information, headways, and other pieces of useful data. Even more interestingly, these feeds often have buried within them the ability to access vehicle GPS data, which provides the ability to find out where a transit agency's vehicles are in near-realtime (limited by the speed at which these location queries can be returned).

Developers have long used feeds published by transit agencies for trip planning applications, which are useful in the extent that they can tell you when you need to be at your stop to get to your destination. However, this information source has always contained within it the potential for much more emancipatory uses. Recall Nate Wessel's guest post here on the shortfalls of Google Transit.

Recently, an innovative new use of this technology has begun to be deployed. In Portland, Seattle, London, San Francisco, and perhaps elsewhere, developers have taken the data provided by transit agencies and done something revolutionary: drawn maps of where all the transit vehicles are at any time.

Obviously the image above is a still from a map whose utility depends upon its continual movement- I encourage you to visit the site to see for yourself. A consistent symbology seems to be emerging, where two colors represent inbound and outbound service – in this case red is going out, and blue in – with separate colors representing the different technologies present on the map. 

Browniefed over at github explains how this all works:

Buses are using short wave radios to transmit there position. They currently send updates every 30 seconds, and TriMet sees the update about 3 seconds later. Each bus gets a turn sending it's position and works like a tolken ring network. Previously it took 90 – 130 seconds and if a bus missed it's turn then it would be double that time (meaning for 1-5 minutes they would not know the position of a bus)

Buses do not actually send GPS coordinates back, what they do send is distance traveled along their trip. They do use GPS but primarily they are using the odometer for tracking position. There is also a door sensor, to verify that a bus is at a stop. There are all sorts of calibration issues with the odometer and drivers have reported that the vehicle is not updating its schedule adherence until the door opens.

So this is cool, but is it useful? How does having access to the realtime position data improve the user relationship with the transit system? These maps don't give any indication of how soon a bus is coming, or how long you will have to wait for it. They don't tell you where the bus is going, beyond showing the position of the next bus up the line. Of course there is no way to determine the time interval between your stop and the position of the bus icon. As a trip planning tool, live bus maps are next to useless.

However, as a method of communicating the shape of the entire network, live bus mapping does something very interesting. It translates frequency into a very simple unit- number of buses. A map reader can see very easily which parts of the city have a lot of buses, and which have few; which places have overlapping services or an interlocking grid, and which are provided only infrequent lifeline routes. Perhaps this technology could someday be integrated or displayed alongside isochrone visualizations as part of a map of the freedom of mobility afforded by a system, but to my eyes, for the moment this is a very technically accomplished curiosity.