Human Transit

was that u.s. news 10-best-transit cities list based on anything?

In the last post, I noted a ranking of the "10 best US cities for transit" in U.S. News and World Report back in February, and some incoherence in how the ranking was explained.  Since then, I've become even more disturbed by the rankings.

It turns out that Michael Andersen of Portland Afoot had done some research, or attempted to:

I was pretty interested in this, in part because Portland's TriMet is, understandably, promoting the ranking heavily. I thought it was great news and wanted to include more details in our magazine.

So I called [Danielle] Kurtzleben, the reporter who'd compiled it, to ask about her data sources and methodology. After five emails/tweets/phone calls over several weeks, I finally got her on the phone, at which point she said she couldn't remember exactly how she'd figured the data, except that it came from APTA and the NTD and that it was "very simple." I asked if she could email me the spreadsheet; she referred me to her editor, who said, bewilderingly, that U.S. News policy is to not share the data it gathers.

So I tried to retrace Kurtzleben's steps. Here's the result; my summary is at the bottom of that page. I pulled the data apart six ways from Sunday, based on the somewhat sketchy description in her article, but couldn't come up with any scenario that ranked anybody above New York City, whose ridership and funding ranks dwarf all others.

Her three metrics were total spending per capita, boarding-rides per capita and safety incidents per boarding-ride. It's not clear what types of "safety incident" counted or how many years of them she analyzed; how she weighted the three metrics into a single ranking. I also suspect she may not have noticed or considered that population data in the APTA handbook is based on population figures from 2000 -- the only place to get apples-to-apples population figures is the ACS [American Community Survey], which she didn't mention using. But even after I ran several variations using the 2000 figures I couldn't duplicate her findings.

Another possibility is that she could have failed to fully account for all the spending and ridership at metro areas that have multiple transit agencies; Portland's relative lack of overlapping suburban agencies would help explain its good ranking. Or she might have calculated population by city rather than metro area.

At any rate, I think I made a good-faith effort to explain these numbers and couldn't.

My one-sentence summary: This article cited out-of-date population figures and was calculated with a methodology that U.S. News refused to explain, based on figures that U.S. News refused to share.

This is a little distressing, especially for a study that's being widely cited by the transit agencies in question.  If you know anyone who might be able to confirm that the rankings are based on, well, something, or anyone else who's tried to do a similar analysis, please send them a link.

Bravo to Michael for expending all this effort in the search of reality.  I don't know if he's right, but he certainly deserves an answer.

More on this topic here!

best u.s. cities for transit?

US News and World Report claims to have identified the 10 best US cities for public transit:

1. Portland, OR

2. Salt Lake City

3. New York

4. Boston

5. Minneapolis-St. Paul, MN

6. San Francisco

7. Los Angeles

8. Honolulu

9 (tie). Denver

9 (tie). Austin

All fine cities.  The methodology:

The rankings take into account per capita spending on public transportation, number of safety incidents per million trips, and the number of trips taken per capita.

But then there was this:

Analysis of data from the Federal Transit Administration and APTA shows which cities are among the best in the country for public transportation. All of these cities' systems have unique features that set them apart. Portland's public transit provides riders with a variety of travel options, including buses, light rail, commuter rail, streetcars, and an aerial tram.

Aargh!  Diversity of technologies says nothing whatever about travel options!  And if Portlanders really did have the options of a bus, a light rail train, a commuter rail train, a streetcar, and an aerial tram all competing for the same trip, that would be a pretty silly network, wouldn't it?

UPDATE:  Followup post is here!

45th century transit information (quote of the week)

I'm unsuccessfully trying to go to sleep reading Iain M. Banks's science fiction novel Matter.  His heroine is thinking of making a trip:

Even without consciously thinking about it, she was there with a diagrammatic ... representation of this section of the galaxy.  The stars were shown as exaggerated points of their true colour, their solar systems implied in log-scaled plunge-foci and their civilisational flavour defined by musical note groups ...

An overlay showed the course schedules of all relevant ships and a choice of routes was already laid out for her, colour-coded in order of speed, strand thickness standing for ship size and schedule certainty shown by hue intensity, with comfort and general amenability characterised as sets of smells.  Patterns on the strands -- making them look braided, like rope -- indicated to whom the ships belonged.

-- Iain M. Banks, Matter, p 95

If you've ever tried to make a clear and informative transit map, look up at the night sky!  Maybe somewhere out there, right now, unimaginable aliens are debating whether "schedule certainty" should be a hue, a pattern, or a smell.

 

 

comment of the week: "expertise as a human being"

Danny sees the deeper theme in my post on the difficulty of figuring out how far people are willing to walk to transit.

I have run into analysis situations like this before. After factoring in psychological bias, demographic trends, purchasing behavior, opportunity costs, and incentives, you end up with an answer that is so complex that it isn't even worth mentioning to anybody. And then everything changes the next month and your analysis is worthless again.

And of course, everybody out there has a theory backed by their own expertise as a human being. And everybody is qualified to say what makes them choose one or another usage pattern. Unfortunately, we tend to extrapolate our own preferences onto others, even though we rarely actually have the same preferences as others.

Randomized controlled trials will work wonders, but for that to happen, people have to be open to the possibility that their favorite theory is wrong. Unfortunately people rarely do things like that...and when they do accept that their theory is wrong, they only do it implicitly after being forced into accepting some other theory by competitive pressure. Its a sad state of affairs.

I'm always struck by how often even highly educated people explain their view of a transit issue by reference to their own experience, as though everyone experiences things the way they do.  Just the other night, for example, an accomplished architecture professor told me that she would ride trains but would never ride a bus.  She preceded this by emphasizing that she knew nothing about transit except what she experiences as a customer, which I later realized was maybe a subconscious way of claiming to speak for all people at that level of expertise -- clearly the majority in most cities.

Although she would never claim to speak for anyone's experience but her own, she presumed she was part of some larger consensus on this question, which made her experience possibly relevant as a basis for public policy.  Watching the larger mass of transit debates, it's always striking how quickly "I would never ride a bus" turns into an unverified claim that "most people would never ride buses."  Most of us want to feel that we're part of the majority, however invisible or repressed.  At another point on the spectrum, you'll hear the same pattern, "I feel x, therefore most people feel x," in claims that transit is an effete distraction from real people's needs because real people (like the speaker) want to drive their cars. 

So I can share all of Danny's reasons for declaring it a "sad state of affairs."  Still, we all get out of bed in the morning, despite it all ...

basics: walking distance to transit

The question of walking distance in transit is much bigger than it seems.  A huge range of consequential decisions -- including stop spacing, network structure, travel time, reliability standards, frequency and even mode choice -- depend on assumptions about how far customers will be willing to walk.  The same issue also governs the amount of money an agency will have to spend on predictably low-ridership services that exist purely for social-service or "equity" reasons.

Yesterday I received an email asking about how walking distance standards vary around the world.  I don't know the whole world, but in the countries I've worked in (US, Canada, Australia, New Zealand) the view is pretty consistent:

• If you have to choose a single walking distance standard for all situations, the most commonly cited standard is 400m or 1/4 mi.  Europe tends to be comfortable with slightly longer distances.
  
  
• However, people walk further to faster services.  (Rail advocates are more likely to phrase this as "people walk further to rail".)  This doesn't have to be a sociological or humanistic debate, though urbanists often frame it that way.  If you are a rational and informed actor seeking to minimize travel time, it often makes sense to walk more than 400m to a rapid transit station rather than wait for a bus to cover such a short distance. 
  
  
• Although the common standard is 400m or 1/4 mi, we all know that this cannot possibly be a hard boundary.  It makes no sense to assume that if you live 395m from a bus stop you'll be totally happy to walk that distance while your neighbor, who lives 405m from the same stop, will be totally unwilling to.  Obviously, the relationship between distance and willingness to walk is a continuous curve without sharp breaks.  This has to be said because our language often forces us to create the illusion of sharp breaks, e.g. when we say something like "people are generally willing to walk up to 400m to transit."

Finally, it's remarkably hard to sift data into a form that produces unequivocal guidance on the question.  For example, the leading US guide on transit planning, the Transit Capacity and Quality of Service Manual, offers only this:

(Source: TCQSM Chapter 3, Appendix A, p. 3-93.  Discussion and version in US units is on p. 3-9.)

This survey-based graph shows the breakdown of local bus passengers by the distance they walked to get to the service.  As you'd expect, few people walk more than 200m in downtown Washington, DC because in such a densely served area, few people would need to.  In low-density Calgary, at the opposite extreme, many people have to walk fairly long distances.

But extrapolating opinions from behavior is a tricky business.  It's hard to reason from how far people walk to how far they're willing to walk.  To do that, you'd have to determine whether each rider would be willing to walk further than he actually has to walk.   You'd also have to speculate about each rider's available options.  If 1/10 of Calgary's bus riders walk 600m or more, does that mean they're willing to?  Or does it mean that these people are so lacking in good alternatives that they feel forced to walk that far?  (The difference between "high income" and "low income" Washington DC suggests that range of options does have something to do with it.)  Sociologists and demographers can have a field day parsing this question, but they're unlikely to come up with an answer of such statistical certainty that it definitively sweeps the question aside.

So we approximate.  We generally assume that 400m is a rough upper bound for slow local-stop service, and that for rapid-transit (usually rail) we can expect people to walk up to 1000m or so.

But when we try to apply these rules of thumb, we hit another hard issue (or at least we do if we're willing to acknowledge it).  Are we talking about true walking distance, or just air distance?  Over and over, in transit studies, you'll see circles around bus stops being used to indicate the potential market area, as though everyone within 400m air distance is within 400m walk distance.

Remember this graphic?

In both images, the red dot is a transit stop and the red circle is an air-distance radius.  If you draw 400m circles around stops based on the assumption of a 400m walking distance, you're implying that the whole circle is within walking distance.  In fact, even with the near-perfect pedestrian grid in the right-hand image, the area within 400m walk (outline in blue) is only 64% of the red air-distance circle.  With an obstructed suburban network like the left-hand image, it can be less than 30%.

Obviously, the market area around each stop should really be defined by the walkable area, which requires a knowledge of the local pedestrian network.  That requires a complete GIS database of every walkable link in the community -- an extremely detailed task that few jurisdictions have been willing to attempt until recently.  Even in Canberra, Australia, which is known in the business for the extreme richness of off-street pedestrian connections, no reliable database of them was available for modelling purposes as recently as last year.

Still, if you don't have such a database, it should be easy to adjust the walking distance standard to reflect the problem.  If you know you have a good street grid, then you can just adjust the radius to reflect the area within 400m walk.  In the right-hand image above, do the math and you'll figure out that if the radius of the red circle is 400m, then a circle whose area is the same as that of the blue diamond -- the actual area in walking distance -- would have a radius of 319m.  So if you want to roughly model the actual radius that arises from a 400m walking distance, and you have a well-connected street grid, draw a circle 319m in radius.  That doesn't give you the correct boundaries of the area -- it's a circle rather than a square -- but it's a far better approximation than just drawing a 400m circle.   I have never actually seen this done, and I'm not sure why.

One reason might be that secretly, we transit planners all want people to walk further.  After all, most transit planners don't want to just passively respond to current behavior.  If they did, they'd all be highway engineers.  Most transit planners believe in the importance of shifting behavior in more sustainable directions, and see both transit ridership and walking as deserving encouragement through intervention.  They are also aware of the public health benefits of walking.

But we have a more vivid motive to encourage walking.  The nature of the transit product is such that if we could stop less often, assuming longer walk distances, we could achieve both better running times and reduced operating cost; the latter could be reinvested as higher frequency.  So the two most fundamental determinants of transit travel time -- running time and frequency -- both depend on our assumption about walking distance.

With such basic things at stake, it's understandable that planners are always looking for ways to push walking distance wider.  That may be the real reason that generations of planners have chosen to approximate a 400m walk with a 400m circle, even though every pedestrian knows how absurd that is.

I prefer to just have the argument in simpler terms.  In Canberra, we pushed the walking distance standard from 400m to 500m, not because people were calling us demanding to walk further, but rather because we looked at how much more frequency and speed we would achieve, and the ridership that could attract, and decided that 100m of radius was a small price to pay for such benefits.  It comes back to that graph near the top of this post, showing how far people walk to transit in different cities.  There's no definitive authority for a 400m standard as opposed to 300m or 500m or even 600m.  Yes, if you pick a bigger radius you'll lose riders from the outer edges of the radius, but on the other hand, you may buy so much travel time and frequency that your ridership goes up.  As with everying else in transit, it depends on what you're trying to do.

can we define "livable and lovable" cities?

That's the nice slogan from a new Phillips Corporation initiative praised today in the Atlantic by NRDC's Kaid Benfield.  The Phillips think tank suggests that we can gather all the qualities of a "livable and lovable" city into three virtues: 

• Resilience, which replaces the more bureaucratic and depressing word sustainability, but means roughly the same thing.  Some great work has already been done on the concept of resilience.  There's already a Resilient Cities movement, and an excellent book on Resilience Thinking. 
  
  
• Inclusiveness, which is about "social integration and cohesion," demonstrated for example in the lack of discrimination or social exclusion based on race, religion, age, and all the other usual categories.
  
  
• Authenticity, which means "the ability to maintain the local character of the city," including "heritage, culture, and environment." 

Below is their graphic summary.  (The PDF [Download] is much sharper!)  Below that is a bit of affectionate heckling from me.

Personally, I have some practical discomfort with the framing of the Inclusiveness category because it is easily exaggerated into visions of a socialist paradise in which we have abolished competition.  When Philips says that "inhabitants should have equal opportunities to participate in the activities of the city," does that mean that when our city's team in the playoffs, we'll give out tickets by lottery rather than selling them, in order to avoid discriminating against the poor?  If we're talking only about nondiscrimination by extraneous demographic categories, fine.  But when you imply that you can neutralize the impact of differences in wealth, you lose so much of the politicial audience -- at least in North America, Australia, and the UK -- that you've probably lost the game.  This issue comes up often in transit, of course, notably whenever anyone suggests that in a capitalist economy, it's foolish not to use pricing to help citizens understand the intrinsic cost of things that they take for granted.  It's a tough one.

Note, also, the lingering contradictory message in their framing of resilience.  On the one hand, the train station signifies that resilient cities acknowledge their "interdependence" with other cities.  On the other hand, the emphasis on local farms and local energy generation suggests the opposite, that resilient cities aspire to greater and greater self-reliance.  This is philosophically interesting, especially because high volumes of international trade -- including in food, which is the opposite of local self-reliance -- are the most reliable mechanism that human society has found to prevent large-scale wars. 

I make both of these comments in the spirit of meditation.  I am not claiming to know how better to define inclusiveness or resilience.  Rather, I'm just marvelling at how difficult it is. 

 

the "cities vs suburbs" trope

We all have too much to read, so here's a tip to save time.  Whenever any article (such as this one) cites information about incorporated US cities as a basis for any claim about trends in the culture, quit reading.  US big-city boundaries are irrelevant to most people's lives, and to anything else that matters about our culture, economy, or destiny. 

Christopher Leinberger makes this point in a New Republic article recently, usefully expanded on by Sarah Goodyear at Grist.  Leinberger argues that "city" and "suburb" is no longer a useful opposition, and that what really matters are walkable urban places vs drivable suburban ones.  True enough, but replacing city with it's near-synonym urban doesn't take us far.   "City" and "suburb" are rich, evocative, and succinct words.  The word city in particular must be fought for, redefined in ways that defend its profound cultural heritage.  The word has an ancient and clear lineage from Latin, one that forms the basis for the word citizen, not to mention civic and civilization.

Greek and Roman political theory was all about the city, in a sense of that word that we can recognize today: groups of people living together in a small space for reasons of security and economy, but also the  site of humanity's cultural and intellectual development.  City is a word of enormous evocative power to capture a range of ideas that drive urbanism.  Leinberger himself can't describe what really matters without using the word urban, which evokes a similar history and resonance.

What Leinberger is really complaining about are discussions of data about incorporated US cities, which are a very narrow and specific problem.  A few of the oldest US cities (San Francisco, St. Louis) have coherent boundaries that describe real cultural and demographic units, but many are bizarre shapes of purely historical interest.

Nobody who understands the lived experience of Los Angeles would claim that the City of Los Angeles is a useful or interesting demographic unit.  While the city excludes a great deal of dense inner-city fabric close to downtown, it has long balloonlike tentacles extending north to take in the whole San Fernando Valley and also south to grab the port of San Pedro.  It also contains a good deal of near-wilderness in the Santa Monica Mountains.

The tentacular, pockmarked, pulsating blob that we call the City of Los Angeles is the map of a long-ago war over water and power.  The only people who care about it today are those who work for city government or serve as its elected officials, plus a few who've considered city taxes and services as a reason to locate in the city or out of it.

Americans should notice, too, that bizarre and misleading city boundaries are largely a US phenomenon.  Europe, Australia, and New Zealand generally allow central (state or national) governments to draw the boundaries of their local governments, so these boundaries usually (not always) end up making some kind of sense.  (With the exception of Queensland, Australian local government areas are too small to have much influence, but that's a different problem.)

As Leinberger says, we need a distinction between walkable urban communities and drivable suburban ones, and American city limits are useless for understanding that distinction.  But the word city -- whose Latin ancestor meant "walkable urban" for millennia until about 1950 -- is still worth fighting for.  Legal US "city limits" are an imperfect and aspirational approximation of what cities really are, and what they really mean for the human project.  Despite their pedantic misuse by the likes of Cox and Kotkin, city limits have no authority to tell us what a city is, and why we should want to live in a real city or not.  The deep attractions and repulsions that we feel for big cities are the key to a longer and truer cultural understanding of what cities are, and of why the civic is the root of civilization.

human transit, age two

The second birthday of this blog passed unnoticed, especially by me, about two weeks back.  I was jetlagged at the time, so confusion about dates was to be expected.

Things have been a little quiet here as I've moved through a series of location changes while focusing my spare time on the book.  But I'm encouraged by the stats. 

At the end of the first year, a year ago, I had written 242 posts, logged 3666 acceptable comments, and about 2000 pageviews per weekday.  Today, those numbers are 585 posts, 9807 acceptable comments, and still in the range of 2000-3000 pageviews per weekday, spiking unpredictably now and then but also crashing predictably every weekend.  So while readership is rather stable, engagement with the material (at least as measured by comments) is accelerating.

The two posts that have gathered the most attention, in links and other citations, are:

• Streetcars: an Inconvenient Truth.  This July 2009 post said something very narrow and factual about the North American streetcar revival movement. 
• Is Speed Obsolete?  The beginning of my debate with Prof. Patrick Condon about the merits of slower vs faster transit services.  Now that he and I are in the same city, we may get to do more of this in person ...

Both of these are the opening of long conversations that extend through several linked posts.  Both also feature rich and interesting comment strings. 

I'm relieved to say that my book (out this northern fall) will steer firmly away from all the technology wars, though the question of transit's ideal speed, for both efficiency and urban form, does figure prominently.  The nucleus of some chapters is already here in my Basics series of posts. 

Per Google Analytics, the total traffic for the last year was:

752,705 pageviews
357,899 visits
116,143 unique visitors

And everyone loves lists of cities, so the top metro areas in readership for the second year were as follows.  The number is individual visits.

24910 Vancouver
18715 Seattle
13006 Portland
11983 Los Angeles
10305 San Francisco (Bay Area)
10229 Toronto
10171 Washington
9628 New York
8112 Sydney
7802 Melbourne
6312 Brisbane
6136 Chicago
4040 London
3434 Minneapolis-St. Paul
2895 Atlanta
2857 Canberra
2821 Auckland

On a per capita basis that's a pretty spectacular result from Canberra (metro area pop. 400,000).  Of course, Sydney, Canberra, and Vancouver are the three cities I've actually lived in during the past year. 

The same data broken down by country, for countries with at least 1% of the total:

212,868 USA
68,025 Canada
27,890 Australia
9,275 UK
4,573 New Zealand
3,776 Germany
3,072 France

... the rest mostly smaller European countries and a thin scattering elsewhere in the world.

Thanks to everyone who's been part of this great conversation so far!  This year will bring a number of changes for me, but I'll do my best to keep this going in some form.  And remember, good guest posts are welcome!

"terror" and "nuclear" options for transportation demand management (comment of the week)

On yesterday's post about the removal of Sydney's M2 bus lanes for construction, mysterious commenter "Quasimodal" laid out a useful theory of Transportation Demand Management (TDM) for such situations:

It's silly, and a missed opportunity, not to provide a bus priority lane through a construction area if you can, even if it takes away from general purpose capacity. What a great opportunity to perform well for an audience that would not usually take transit!

But there is a technique I call the transportation demand management [TDM] campaign of terror, which is fantastically effective at reducing construction impacts. If you blow a big enough horn about how gawd-awful the traffic will be during construction, that traffic will almost always not occur. People will rearrange their vacations and do whatever they need to do to avoid the problem area.

I don't know how long this [Sydney M2] construction is supposed to continue, and the TDM campaign of terror can't persist indefinitely, but if it's only a few weeks it's possible that the buses rerouted to local arterials will be a lot slower than those that brave the highway traffic if that traffic doesn't show up!

Note that the TDM campaign of terror is different from the TDM nuclear accident, in which one destroys one's economy, causing plummeting employment and related travel. That strategy has been very effectively applied over the past couple of years, reducing traffic volumes all over the world (with the unfortunate side effects of homelessness, despair and civil strife...) Fortunately, it seems governments around the world are establishing policies to continue this strategy for the foreseeable future, using anti-growth austerity programs. As a planner, I'm happy to be freed from the the endless cycle of growth, and to focus on a more environmentally sound steady state (though probably unemployed) future.

 

 

u.s. "transit lovers unite"

A new campaign by the American Public Transporation Association:

Washington, DC –The American Public Transportation Association (APTA) and its more than 1,500 members are launching “I <3 Transit”(I heart transit), a mobile text campaign, to engage riders across the country  and demonstrate the importance of public transportation to individuals, communities and the nation.  The campaign, which starts this week,  allows participants to show their support for increased public transit funding by simply texting “transit” to 86677.

Each year 10.2 billion trips are taken on the nation’s public transportation systems.  Riders depend on their local transit provider for freedom, mobility and access.  By texting “transit” to 86677 public transit riders and enthusiasts alike can stand together in support of public transportation.  This simple action helps spread the message that Americans want public transportation as a travel option in their daily lives. 

Supporters can also follow the campaign on twitter at @ihearttransit.