« social media's influence on public transit (guest post) | Main | graphic artists! seeking ideas for book's cover »


Feed You can follow this conversation by subscribing to the comment feed for this post.

Corey Burger

One-way loops are a good idea at outer low-density ends of lines,

Why is this true? It is because you can cover more ground?

Great post, btw. Especially the point about the details.

Jarrett at HumanTransit.org

@Corey Yes, because in low-density areas all you can do is cover area to provide basic links to stronger transit services.  One way loops often do that well, though even there you can lose riders if the "long way around" the loop is too long.

Eric O

This pretty much applies, almost verbatim, to Charleston, SC's geography and destination conditions. But if you want to see a system philosophy that appears to be "one-way loop" by default just take a gander at Charleston's bus route map. (http://www.ridecarta.com/shared/pdf/System_Wide_Map.pdf) Confusing as all get out!

Of course, here, there are special conditions that disadvantage a true high-frequency network. Narrow (slow) streets in a core that is largely touristic (horse buggies are your alternate transport), fewer chokepoints that discourage a dispersed network, and a much longer N-S peninsula market (demanding a more linear service trunk). But Charleston's core could be a good case of comparison with Halifax (I'd be interested to see how they compare at the region level too).


Looks mostly good. However, I don't like the fact that the dark blue line turns right to go into downtown, while the purple line is the one to go across the bridge. This creates a situation where traveling in a straight line leads to a forced transfer, in favor of a direct connection for an l-shaped trip to a preferred destination. This smells like a radial, rather than a gridded system.

Furthermore, it appears from the map that to get downtown from almost anywhere on the dark blue line, you could do it just as easily via the orange, purple, or light blue. The trip NE from downtown across the bridge, also seems a bit redundant with the ferry. And the walk from downtown to a straight-through blue line still looks to be under a mile, if I read the scale correctly.

So, perhaps this might be better if the dark blue ran straight through across the bridge, while the purple ended downtown, by the ferry dock.

Interestingly enough, as you constructed the network in stages, the purple going over the bridge looked perfectly fine when that was the only line shown. It was when the blue was added that things started to look a little bit strange.


@Eric The question of whether to have a straight east-west line that misses downtown is, I suspect, something that can only be answered with data on ridership patterns. Or else you may end up with a blue bus that completely empties out at the transfer point with passengers who try to cram on to the already-full green buses, only to be filled again with passengers coming from downtown, which seems operationally suboptimal.


Drawing transit routes based on data on ridership patterns has two problems:

1) There are lots of things that can cause ridership patterns to naturally change with time. For example, businesses can expand, go bankrupt, or relocate. Changes to zoning laws can increase or decrease the number of people that can live in a neighborhood. Professional sports teams can move into our out of a city. And changes in schools, crime, etc. can make a neighborhood a more or less desirable place to live.

Because of this, any system that has special kinks in the route to handle the demand optimally for the ridership now is likely to be suboptimal for the ridership ten years from now. While this can conceivably be addressed by redrawing the bus routes every few years, change leads to confusion about what routes go where. Plus, route changes can be politically very difficult to pull off, even if most people are better off because there will inevitably be someone who's worse off, who will complain loudly.

2) Using ridership data is inherently an incomplete picture, as it only includes the people who are riding the bus today, not all the trips that people are making. For example, ridership data may say most people riding the bus are headed downtown, but are most people, overall, really going downtown, or does it just appear that way because almost everyone that isn't going downtown has decided that the bus is too slow and is driving instead? If you want your mode share to grow, rather than merely cater to your existing customer base, these are important questions that need answering. Jarett has a good discussion on this.

What's nice about the grid pattern is you don't have to worry about things like "we have a bunch of people going from this point to this other point". You simply focus on connecting areas according to the geometry of the roads and let whatever travel patterns people want to have happen.

Ridership data can be important, for example, in deciding which size vehicles to use on a given trip. For example, if most people really are headed to downtown, you can use smaller buses on a straight-through dark-blue line, with bigger buses on the pink, light blue, and green, to handle the loads. If ridership patterns change in the future, reallocating which sized buses go on what routes, or adding extra buses on the busiest routes during the peak, is much less disruptive than redrawing the routes themselves. By contrast, trying to use ridership data to draw the routes themselves simply leads to a situation where you're enumerating a handful of trips that you're designing the bus for and, for everything else, people should just drive. This narrows the customer based to a few niche markets, rather than serve a broad base of the population.


If ridership patterns change in 10 years, can't you just change the bus routes to more closely match demand? Isn't that one of the big benefits of the much-vaunted "flexibility" of buses?

And as far as grids go, they have their pluses and minuses, and one of the minuses is that cities are rarely isotropic, with demand evenly spread throughout the grid and trips evenly distributed in all directions. Very often, there will be a center, where more people want to go to, or maybe even a few such centers, and it can make sense to divert buses there.

And as far as serving niches, it can be a choice between providing decent service to a few, and winning just that small ridership market from the car, but decisively, or providing mediocre service to everyone, and not winning against the car at all. Which one you really want depends on the transit agency, its goals, its political environment, and what is physically and politically achievable in a given city.


This reminds me of Madison, which is an isthmus where they've funneled many routes onto two streets through the isthmus - no need for cross-isthmus service when it is one mile wide: http://www.cityofmadison.com/metro/schedules/SystemMaps/WeekdayMap.pdf

Marcus Garnet

As a professional planner and resident of Dartmouth across the harbour from Halifax, I found this really interesting and insightful. That #1 bus is indeed very successful in terms of ridership and convenience! As for "continuing the blue line across the bridge to Dartmouth", a few years ago such a route was in fact added - the #10 bus - and has also been very well received. It also extends in both directions to link our two largest business parks. Last but not least, an underused, grade-separated railway line follows the diagonal southwestern edge of the Halifax Peninsula, links four of the six choke points, and terminates at the southern edge of downtown and the eastern edge of the university corridor, opposite a new farmers' market and arts college campus. In the other direction it extends all the way out to Bedford, a major suburb and shopping destination. Let's "join the dots!"

Jarrett at HumanTransit.org

@Marcus.  Thanks for your local perspective! 

As for the rail corridor, by all means preserve it, but meanwhile, build the market with buses.  Once you have the Line 80 bus to Bedford running every 15 minutes with healthy all-day loads, you're ready to think about light rail, but as I understand that needs some expensive duplication of track.

As for commuter rail, again, it needs to be more cost effective than buses (or ferries) in the corridor.  Because you can deploy three modes (rail, bus, ferry) between downtown and Bedford, you risk seeing all three developed separately, as different projects, stopping at different Bedford stations, and thus all competing with each other to achieve a poorer frequency than could be accomplished with a strong investment in one.  Note too the need to be clear about whether you want to serve a commute-only market or an all-day market.  A few commuter trains on your abandoned line might be possible at lowish cost, but of course only all day high frequency would be useful for intra-peninsula travel. 

Alon Levy

It's a really interesting process, and I'm definitely going to keep referring to this when I finally get around to writing about Tel Aviv transit, especially where its subway should be built.

Alan Howes

Splendid stuff. At every 10 minutes, Line #1 is not bad as a core service. But surely there is a case for extending it beyond its current rather short span, beyond Mumford Terminal and/or Bridge Terminal to key suburban destinations? Or does someone at Halifax Transit think that 40 minutes is long enough for a route, and extensions would prejudice reliability?

But my first thought on looking at that "shaded blocks" map was - Downtown Circulator! Perhaps operating fares-free, perhaps with small buses, certainly every 10 minutes or better, both ways round a circle serving at least four of those five blocks and the Barrington / Duke Ferry Terminal. (That dead-end to St Mary's is a pain.)

Jarrett at HumanTransit.org


Actually, encouraging the permanence of frequent bus network elements is important if you want people to make locational decisions in response to them.  The main lines in this structure arise so directly from permanent elements of the urban form that there's no reason they can't be permanent. 

Re your comment on grids, they actually make your point: they are appropriate to areas where the continuous density is high enough to support decent service for anywhere to anywhere trips, and where there enough major destinations to anchor them.  All of the Halifax peninsula is such a place.  Decisions about whether to serve some people or everyone poorly become much more acute in suburban areas.  Jarrett

Eric Doherty

This is interesting. I have a bias against U-shaped transit lines, yet this example is full of them.

Maybe I am too fond of straight lines and L-shapes, but I find 'going backwards' to get to my destination irritating.

I don't see this as so much of an issue with longer lines with a short hook at the end, as in Vancouver Canada's downtown Expo Line light metro (AKA SkyTrain).

Jarrett at HumanTransit.org

@Eric Doherty.  U-shapes are designed to connect endpoints to the middle, not to connect endpoints to each other. 


Does this analysis factor in the biases in public policy which lead to the creation of the limited choke points in the first place? Notably the unwillingness of the city to take on south end homeowners (as opposed to those on Chebucto and Bayers for instance) by putting corridors via the railway cut, or streets like Oakland road and an extended Connaught Avenue?

Most of the city's traffic and mobility problems are linked to this failure, given the location of core institutions near the south end and the expansion in south end of downtown. No transit and roadway plans will work if these problems are not rectified by the expansion of such south end arterials and construction of south end bridges over the arm, and to circumferential in Dartmouth.

As your maps clearly show, the South end is the key gap in any logical transit and traffic grid. This failure has also lead to dysfunctional spread of housing northward to Hammonds plains, wasted infrastructure capacity on Northwest arm drive and devalued property values in Mainland south and inflated ones in Mainland north and south end, distorting development and limiting cohesiveness of urban space, complicating transit planning completely.


PS the hasty decision to pull up rail lines to Fairmount, Timberlea and beyond cost an opportunity to build in a commuter rail option which could help many bypass the choke points.


Also as the presentations on the Facebook site indicate, the thinking is solely about those who live and move on the Peninsula with little reference to those who do not.


How could the system be extended to those beyond the choke points who have been most disadvantaged by these biases in policy favoring the South end?

Ashley Morton

Two comments here. (Full disclosure - I live in Halifax.)

In advance though, Jarrett, I attended the workshop you participated in here in Halifax, and thought your presentation was really great, given the time you had. It certainly got me thinking about chokepoints and frequency in ways I hadn't before.

My wife and I moved to Halifax 3 months ago, and sought out a house on the Peninsula in order to acheive our desire to go carless. Of note - while it's not necessarily tied to the ability to live without a car, we paid about 20% more for that priviledge when purchasing our house than much of the non-Peninsula area would have cost. On the other hand, we expect to save thousands every year on transportation costs. We live in the North End (pretty much the northernmost point of your bright green line.)

First - thank you for some free work on Halifax's transit questions - I think it makes a great contribution to our conversation!

Second - was there any particular reason why you didn't use Gottingen/Novalea (the longest straight yellow line on your map - the northern half of Gottingen is renamed Novalea)? I realise that it doesn't provide "bounce" capability at the northernmost end (it doesn't connect to those northern-boundary roads that it looks like it should), but it could easily terminate at a large community college campus there at Leeds & Novalea, giving it the traditional "go one way, turn around and go back" aspect many grid-based frequent lines have, with the sought-after "anchor" at one end. Alternately, it could relatively easily go over the northern bridge, something none of your sketched-in lines currently do. A final option would be to have it take over from the #9's (green) job, so as to avoid running two similar buses so close (better to have one, and more frequent) - Gottingen has much more street-level "stuff" - theatre, pubs, restaurants (and, oddly enough, a military barracks, but hey - they generate public transit trips into downtown!), while Barrington has nothing but nice views of the water, and stops at the dockyard and the shipyard - better for a limited-stop route, not a frequent one, in my opinion.

Thanks for a fantastically educational blog. I have started reading your old posts, while moitoring the "Halifax" link on the right. I really appreciate your thoughts.

Jarrett at HumanTransit.org

Ashley. Agree re all your points re Gottingen-Novalea. It needs a
line but it has too many anchoring options, none of them close to
ideal, Note how much easier it would be to integrate into a network if
you had a connection facility at the northern end, which is also a

Ashley Morton

Here's something that would be AWESOME, but I imagine impossible to pull off - build an HOV-only onramp to the bridge at the north end of Gottingen/Novalea. You only have to go about 20 feet! ...Though I can't quite figure out how it could be signaled... Building the *on* to the bridge part would be relatively easy. Not so much the other way.

Are you aware of any HOV-only highway accesses? (and not places where it only serves a garage or something - I'm talking about places where single-driver cars would really like to go, but they've been excluded from doing so...

Jarrett at HumanTransit.org

Ashley.  Seattle has HOV-only access ramps to certain parts of downtown.   I think you'll find these in several cities. 


Awesome post, really helpful and interesting exercise, so to start with: thanks!

A few points on the map: I find it interesting that #9 (light green in your map) goes all the way to Barrington Street before turning down downtown. Now, admittedly I'm not too familiar with Halifax, only having visited, but it doesn't look like a very high-demand corridor right next to the water industries, but rather the frequency of a grid route might be better suited to Novalea and Gottingen. I notice instead Metro Transit runs #3 down that route, which looks like a rather circuitous route. Why not just have the light green line turn early? Any Halegonians who can explain the thinking? (I see I'm not the first person to think this.)

The one thing about your map I will say is I wouldn't ever ride the light blue downtown from anywhere that's not along Inglis. It's too circuitous for that. Still, that's probably not a particular problem, and better than the existing one-way loops.

I also find it strange that there's a bus route running down Cunard (#5) in the real life system, when that could be combined on North to add more frequency and only ever be three blocks away. There's numerous examples of this in the Halifax system.

To be honest, the actual system seems rather more complex than it need be on the whole. It seems every suburban bus that comes into the peninsula takes a slightly different route into downtown or to the universities. This in particular caught my eye: #41 replicates #1 almost exactly, except going down South instead of Spring Garden. But the market served is basically already served by another bus that would then have to run less frequently to make funds for this bus. That seems very strange to me. http://www.halifax.ca/metrotransit/Schedules/images/Route41Rev.gif

Sean Gillis

@ Tessa,

The #3 bus runs every two hours, maybe 4 times a day. It is a special route connecting nursing homes directly to downtown and major shopping centres. It was scheduled to be terminated, but got saved by Halifax Council who were ashamed to take away service from seniors, despite the route being curvy and overlapping lots of other high-quality service.

The current route #9 is a well ridden bus, providing access to the Naval base (thousands of employees, located just south of the southern bridge). I agree with everyone who has pointed out that Gottingen/Novalea would be a better route for a high-frequency corridor. Barrington however still warrants some service, which would be the details Jarrett was talking about.

There is an important historical reason for many of the peninsular bus routes - they follow portions of the old streetcar system. The #1 route (pink), #9 (green) and #7, which runs down Gottingen and isn't shown, are the best examples. These routes do lots of useful things and could have formed the basis for a great network. However, routes have clearly been added that try to serve everywhere to everywhere. The system is on-godly confusing and wasteful. Spring Garden Rd. and Barrington for example currently have dozens of moderately busy buses during rush hour, each running low frequencies. Meanwhile Quinpool Rd., Gottingen, Robie and University are underserved.

A note on the ferries. They are very useful if you are going from Downtown Halifax to Downtown Dartmouth, or vice versa. Two problems arise, only one of which I think is easy to fix. The tougher problem is that the Dartmouth Commons and a golf course present some bizarre geography and sort of cut off downtown Dartmouth from most areas north of the Bridge. I think it's just easier to route people north of the MacDonald Bridge through the Bridge Terminal.

The second problem is on the Halifax side there is NO service between the ferry terminal and the universities, hospitals. I don't joke, there's not even a shuttle. You have to walk a couple blocks uphill to Barrington to make that sort of connection. Halifax does have a new downtown shuttle, but it provides a half hour frequency loop around the old downtown core. I haven't seen the numbers but I imagine it's a failure due to the low frequency and short distance walk trips it competes against. So short term the shuttle would be much more useful connecting the ferry to the huge number of jobs/ students around University. Long term I think this connection (ferry to universities) could be a key cog in a high frequency network. The ferry terminal is in some ways another choke point, and since it also is dead downtown it could warrant a lot service.


That's a good design.


I think Halifax should use the land liberated from the eventual dismantling of the Cogswell interchange to make a new transit hub. One of the piers of Purdy's Warf could be leased or bought and used as a fairy terminal, making it easier for Dartmouthians (or maybe even Bedforders, depending on the future of THAT ferry) to connect to the cities transit. Many lines end near there at Scotia Square anyway, but its just a bit too far from the ferry terminal for an easy change. It's also within walking distance to the Naval Dockyards (albeit depending on where one works within the Dockyards) and, of course, downtown.

Additionally, if Halifax ever WERE to get trams (which, after reading other posts on this site, I'm beginning to think might not be such a good idea), the end of Cogswell would be a prime terminus. A tram line could run all the way up Cogswell to Robie, and all the way up Robie to at least Saint Mary's or even turn down Inglis and end at the Farmers' Market (an area that seems to be growing). For almost the entirety of Cogswell and Robie it could run in the median and thus be much MUCH quicker than personal or even bus transit during rush-hour (which can be surprisingly bad for a city of our size). It would also service the medical corridor and the universities. Any thoughts?

The comments to this entry are closed.

the firm

Jarrett is now in ...

Related Posts Plugin for WordPress, Blogger...
Related Posts Plugin for WordPress, Blogger...