Skip to main content

Simplify and Connect: a key to better bus networks

What if I said your city could have better public transport without more funding or higher fares? Does that sound too good to be true?

Reinventing Transport this time is a basic explainer for the idea that public transport networks are often improved by being simplified. It can be a low-cost step to a better system.  

If people in your city face long waiting times for buses, there might be too many bus lines. A simplified network may offer better service.

Does that sound intriguing? Or maybe this issue is old hat for you. Either way, I hope you will get something from listening to the episode or reading the article below.

The basic idea

Imagine a town with 100 buses. And suppose the town has 25 bus lines. There would be four buses for each line. But suppose the town simplifies its bus network down to just five lines. Now there are 20 buses per bus line.

Does the new network have higher bus frequencies? Of course it does. But the people using the new network now make 'transfers' ('connections') much more often to complete their journeys.

Which is better? People hate waiting for buses. But do they hate connections even more?

These two networks represent exaggerated illustrations of two real responses to this dilemma.

In real life, many cities do indeed try to minimize connections and end up with complicated networks with many lines, each with low-frequency service.

But some cities take a different path and simplify their bus network to get higher bus frequencies. They also work hard to try to minimize the pain of the resulting transfers.

It turns out that we usually CAN come out ahead when we simplify a bus network.

The simpler network allows higher frequencies, meaning shorter waiting times for the first bus in a journey AND shorter waits for any connecting buses along the journey. So if all goes well, most existing journeys will end up taking less time in total, even if they now require a connection.

It turns out that these simpler, connective networks are often more popular than what went before. At least, they are usually popular once they are implemented.

Let's take a look at Barcelona [3:27]

The transit authority in Barcelona, TMB, described the old bus network like this:

“the Barcelona bus network was the heir to the tram network of the previous century. As the city grew, the routes were extended and superimposed until a network was created that was illogical, with repetition of lines and journeys that were slow and buses that were infrequent. Neither were connections ensured between different parts of the city and for the uninitiated user, the network was hard to understand and read on a map.”

Does that sound anything like bus networks where you are?

Barcelona has a significant Metro and suburban rail system but was not getting the best from its buses. Average headways on the old network were around 12 minutes. That may not sound too bad but Barcelona is has the highest urban population density in the western world. Public transport trips there average only 7.2 kilometers and I would assume bus trips would typically be even shorter. Waiting 12 minutes is a lot to go only five kilometers or less.

So Barcelona came up with a plan for its Nova Xarxa - a new network - inspired by modeling by Carlos Daganzo of UC Berkeley.

Now that the new network is actually fully operational, the new network has 28 high-performance lines. There are 17 'vertical' lines running from the sea to the mountains, eight so-called 'horizontal' lines and three diagonals.
Barcelona's Nova Xarxa network of high-performance bus lines. Source: TMB

According to a 2017 paper by Hugo Badia, Juan Argote-Cabanero and Carlos Daganzo, “The Nova Xarxa will eventually be served by 573 buses with an average headway of 6.18 min, similar across all lines. Contrast this with the old bus network which was served by 761 buses with an average headway of 12.30 min. Thus, the Nova Xarxa will use fewer buses but deliver nearly twice the service frequency of the old network.”

On the old network, 13 percent of bus trips involved transfers, which was actually already quite high.

According to that same paper, in 2016 26% of boardings on Nova Xarxa came from transferring passengers and they expected this to rise to 44% once the entire network was deployed.

I should mention that it looks like Barcelona may have not have been quite as ruthless in eliminating old bus routes as originally proposed. Or perhaps they have decided to phase some of them out more slowly.

But it IS clear that they have established a simple grid of high performance bus lines.

Ridership results cited by Eric Goldwyn and Alan Levy look good so far.

Other cities doing this [6:40]

Various cities around the world are doing something like this.

In fact, Canberra, Australia's capital city shifts to its new network next week!

Portland, Oregon in the United States was a pioneer with its simple grid of frequent bus lines adopted in 1982.

Simple, connective public transport networks seem common across central Europe. Zurich is a famous example. Vienna clearly has a grid like network of trams, buses and metro lines as you can see from this map showing part of the network just to the west of the old central area.

Another example is San Francisco's Muni network.

Jarrett Walker and his firm have been helping various cities to redesign their bus networks along these lines. Houston is a prominent example as well as Richmond and several others. Dublin is a live case right now

And Auckland's successful new network was discussed as part of Reinventing Transport Episode 10. By contrast, a similar effort in Wellington seems to have gone somewhat awry.

Santiago de Chile is an example with its TransSantiago reforms. A connective network was part of this but they tried to do many things all at once in a big bang change, which was initially disastrous. I want to follow up to find out how things are going there now.

Bogotá in Colombia has been reforming both the bus network and the regulatory approach to non-BRT buses with mixed results for now apparently.

A success story in Asia is Seoul with its famous 2004 bus system shake up.

I was in Yangon in Myanmar recently and was surprised to learn that Yangon has also been reforming its bus network, simplifying the routes, reducing the number of routes and trying to have a better level of service on each of those routes.

I also understand that Bangalore in India has been piloting some network changes on some of its bus corridors with the help of the WRI Ross Center for Sustainable Cities.

And Jakarta seems to be trying to improve coordination between its informal minibuses and trunk services, including the TransJakarta BRT.

What does a bus network that sorely needs simplifying look like? [9:55]

One kind of problematic bus network focuses too much on the city center.

In cities in the developing world that have unregulated minvan-based public transport for example, it is common for public transport networks to look like a set of bamboo groves with large numbers of parallel stems that then spread out towards the edge of the city.

Seoul's old bus lines also looked like a set of bamboo groves before the 2004 reforms.

A similar thing seems to have happened in a lot of relatively car-dependent Western cities (although most of these have regulated formal-sector bus systems). Their bus networks, like Barcelona's old one, are often based on historic lines from 50 or 100 years ago. They typically haven't been updated for their new situation as bigger modern cities with multiple destinations. Many have given up trying. So most of the network is focused on radial service into and out of the core area and mainly on peak hour commuting.

Another kind of bus network that needs simplifying can be found in cities with more successful public transport. Big multimodal cities like Barcelona or Singapore usually don't have a bus network that is too focused on the city center (although Barcelona did have a little bit of that problem with its old network). Public transport in these cities is more ambitious. In fact, the problem is they have been trying too hard to provide 'ubiquitous networks' by trying to provide too many lines to too many places from too many places. So each of those lines ends up being infrequent.

The resulting bus networks end up looking like tangles of wires.

You can see this on the left hand side of the title image at the top of this article, which shows a small part of Singapore's bus network. And in Singapore, many bus lines have waiting times throughout the day between about 12 and 15 minutes and sometimes up to 20 minutes. Like Barcelona's old network with 12 minute headways, that may not sound bad. But it is not good enough in a dense transit-oriented city like Singapore where the average bus trip is only 5 km in length.

And if cities with this kind of 'tangle of wire' network try to get better frequencies by simply throwing money at the bus system they would risk nasty bus congestion problems because of all the overlaps in the network, where numerous different lines run on the same sections of road.

So let's take stock.

Your city probably needs a simpler, more connective, bus network if many of your bus lines have fewer than one bus every 15 minutes. In fact, in dense cities, having a bus less than every 10 minutes or so is probably not good enough.

You probably need a more connective network if your bus network is so complicated that you can't put it on a map easily or if the map looks like a tangle of wire.

Your city probably needs a more connective bus network if the bus network looks like a series of bamboo groves sprouting from the city center, but with almost no lines in any other direction.

[Keep reading below]

IF YOU LIKE THIS, please share or recommend Reinventing Transport to anyone who might be interested. And subscribe, if you haven't already (it's free):
  • sign up to get updates by email from this site
  • subscribe to the audio podcast (search for 'Reinventing Transport' in your podcast player app or click the symbol that looks like a wifi icon in one of the players at the top or bottom of this article).
You can help me make time to continue this work by becoming a Patreon patron.

Who are some key people on this?

I already mentioned Jarrett Walker. His Human Transit blog and book are indispensable in highlighting that this is an opportunity that communities may want to seize (depending on local goals and priorities).

There have also been many other champions of simpler, more transfer-based public transport networks. They include numerous practical people behind the scenes in many cities. For example, in Portland, Jarrett Walker mentions Ken Zatarain and Thomas G. Matoff as especially crucial for the 1982 bus network changes.

The late Paul Mees from Melbourne in Australia was influential. Paul was inspired in part by the PhD thesis of my friend Felix Laube, who is from Zurich. That thesis included a clear explanation of Zurich's connective approach, with a frequent mesh of lines in the core and a sparser, timed-pulse approach to enable convenient connections even in outer areas with lower-frequency service.

Gustav Nielsen from Norway has been another prominent proponent of this kind of network reform (pdf).

And I already mentioned Carlos Daganzo.

So what's the catch? Gaining enough support can be difficult! [17:50] 

Simplifying a bus network can provide rewards for a small investment but this can be a challenging reform to enact. Implementation needs to be carried out well and sometimes goes wrong. But even before that, getting any kind of consensus to proceed is always difficult.

Some people currently using the bus network will feel they are losing the bus lines they are used to.

Many will need to make a transfer in their daily routine that they didn't previously have to make.

We can explain that most people are going to save time and expand the range of reachable destinations but many will care more about the specifics of what they are losing.

If these perceived losses are concentrated in certain neighborhoods or among clearly defined groups of people, then organized opposition becomes likely.

It doesn't help that there is usually also confusion and sometimes even a disinformation campaign from opponents.

Dublin's ongoing 'Bus Connects' efforts to remake its bus network have also seen opponents assuming that the network changes were all or nothing and that the consultation process must therefore be a sham.

Bus network simplifications are also a great example of the fact that revealed preference is more reliable than stated preference. In other words, we can trust how people actually respond in practice to a change more than what they say they will do.

So it is important to design a process that helps people see key trade offs and face up to their consequences. Here is an example of Jarrett Walker laying out the key trade-offs for Dublin very early in the Bus Connects process so that the subsequent network design exercise would be well informed by local priorities.

Another problem is that people's current experience with making a connection between bus lines may be abysmal, with unpredictable and long waiting times. That experience can make it difficult for people to imagine making convenient bus to bus connections in a future simplified network.

So embarking on a bus network reform is not easy. It's not for the faint-hearted, but the rewards are great.

If you're interested in this issue, I really recommend visiting Jarrett Walker's Human Transit website and looking at network redesign stories from various cities.

Despite the difficulties, this approach seems to be gathering some steam. It is also making a tangible difference in many cases. For example, in the United States bus ridership has been declining recently in most cities but those few that have redesigned and simplified their bus networks seem to be holding up much better than the others.

Let me end by restating some key points [23:48]

If your bus network has lots of lines that overlap each other, if it has low bus frequencies throughout most of the day, then your city could probably benefit from this kind of bus network redesign. 

The key things that a transfer based network or connective network needs for success are a simple and easy to understand network with easy transfers and high frequency services. 

It's not easy to do this, but the rewards seem to be great and they seem to be lasting. 

Cities with successful public transport around the world tend to have rather connective public transport networks. 

There are variations on this that suit small towns and big cities, there seemed to be variations that suit dense transit oriented cities and car oriented low density cities.

Is this something that you could be championing in your own city or town?

Video created with Headliner.   Writing this article was aided by a transcript created by


Popular posts from this blog

Podcasts on urban mobility and urban issues: a LONG list

Below is my increasingly long list of podcasts on urban mobility and/or urban issues. 

If you are not yet a regular podcast listener, you need to download a podcast-listening app to your phone, tablet or desktop and subscribe (it's free) to the podcasts that interest you.

UPDATE 1: This list has many podcasts but obviously I hope you will try mine! They are Reinventing Transport and Reinventing Parking.

UPDATE 2: I have added THIRTY NINE more since this was first published. Thanks to everyone who has sent tips.

Transport-based City Types and their Trajectories

I want to help you get perspective on your city and its transport system with the help of simple city types based on their dominant transport modes, such as Walking Cities, Transit Cities, Bus Cities, Motorcycle Cities and Car Cities.

This way of thinking about cities is a heuristic (an imperfect mental model or technique that is nevertheless good enough to be helpful). And it obviously is imperfect. For example, real cities often have various modes of transport, and modern cities are really all some kind of hybrid city type.

But it is still useful, especially if we add the idea of a Traffic Saturated City, which is a very different beast from a Car City. It is important for change-makers in Traffic Saturated Cities to be aware they are not in automobile dependent cities yet.

Options for digesting this: 
Read the brief article below and study the diagrams. They complement the podcast. For more depth, LISTEN to the 37 minute audio with the player above. A full transcript of the podcast is…

Singapore Urban Transport: The Warts-and-All Story

Singapore's National Day is this week (9 August). So I decided to share Singapore's urban transport story - or my slightly unusual take on itIt isa unique city in various ways but its urban transport policies are well worth your attention even if you don't live in Singapore.

This is a warts-and-all version of the story, and it is my own view, not any kind of official one.

It's also a little wonkish in parts. [Hi all you policy wonks!]

But I hope to keep your interest with some surprising twists, such as:
Why was the bus-only public transport system in an awful state by the early 1970s?If the buses were awful in early 1974, how was Singapore able to impose drastic increases to the cost of motoring in 1975?You will have guessed that the buses must have been drastically improved in 1974/75. But how was that achieved?Singapore urban transport enjoyed success through the 1980s and 1990s but its core social bargain (cars for the rich; decent but basic public transport for …