Why Do Buses Keep Bunching?

January 25, 2013 at 8:24 pm

It’s exasperating and anyone who even occasionally rides the bus knows the feeling. You’ve been waiting for the bus for at least ten minutes even though Google Maps told you it was supposed to come five minutes ago. In the distance, you finally see the twinkling lights that look like the bus. The bus pulls up to the stop, and right behind it is another bus, on the same route, going to the same destination. A third bus might even race up behind these two. The buses are bunched.

Surely, if the buses were running at the proper headway, you’d have been on one a few minutes before, and two or more buses going to the same place wouldn’t be leapfrogging each other. If you’re lucky, bunching is just a rare occurrence that comes with the low price of using public transportation. When its more than just an occurrence, it becomes a problem that can undermine a bus route.

Screen showing bus times for the northbound 49 bus at 6, 6, 6, and 7 minutes.

Sometimes, Buster (an app for CTA bus and train arrival times) shows the same bus time twice. But not here. 4 buses are approaching, all at the same time.

I’ve written before about my frustrations with the #22 Clark bus in Chicago. Quick background: The #22 bus constantly bunches and is one of the most-boarded bus routes in Chicago. It’s an unreliable route (I’ve waited as long as 40 minutes–during the day–for a #22 bus). I obtained bunching data from the CTA and have uploaded it to Google. Feel free to play with the data yourself. Here’s what I found:

10 Most Bunched CTA Bus Routes, January – October 2012


Bunch rate*

Bunch rate, 2011

Change from 2011

Total Rides

% of total bus rides**

System Rank†

66 6.56% 5.40% 1.16%  7,247,542 2.75%  4
49 6.48% 6.10% 0.38%  7,786,798 2.95%  3
134 6.19% 6.80% -0.61%  699,702 0.27%  83
79 6.07% 6.00% 0.07%  8,574,625 3.25%  2
3 5.93% 5.50% 0.43%  6,054,093 2.30%  8
22 5.87% 4.50% 1.37%  6,253,704 2.37%  7
135 5.80% 5.60% 0.20%  849,017 0.32%  78
151 5.45% 5.20% 0.25%  5,953,055 2.26%  9
147 5.13% 4.40% 0.73%  4,288,372 1.63%  18
20 4.85% 5.80% -0.95%  5,441,406 2.06%  13
*Bunch rate is % of timepoint headways less than one minute for buses running the same route/direction, Jan-Oct 2012
**% of total bus rides is the total number of riders for each route divided by the total number of bus riders, Jan-Oct 2012
†System rank is the rank of the bus route in terms of total boardings among all CTA bus routes, Jan-Oct 2012

See all data: CTA bus boardings by month and day type, 2012 | CTA bus bunching by route, 2010, 2011, Jan-Oct 2012

The percentage of buses bunching is the amount of buses (on the same route) at certain timed headways that are less than one minute apart (going in the same direction). I’m not sure if the location of the timepoints is available. If the timepoint location information were available, a map could be created to find out where the bus bunching happens, how long it lasts, etc. I wonder if this could be made by creating an application that pings bus tracker locations every minute or two and records the locations of the 20-or-so most bunched routes. Maybe someday.

The CTA recently released an FAQ of sorts on its website called When Things Go Wrong, and bus bunching is on the list. After explaining why bunching happens, they suggest remedies, but they’re not that great (and they acknowledge this):

  • Running buses express/short-turning buses: This doesn’t work because buses don’t have signal priority and mix with private vehicles. You can’t run a bus “express” unless the street is empty and every light is green. Short-turning a bus (putting it out of service and turning it around) also isn’t attractive, because it forces riders off the bus and onto another (presumably crowded) bus.
  • Introducing “wiggle room” or adding service: Also not the best remedies, because it either slows down service purposefully or creates so much service that bus travel might not even be environmentally friendly (too few passengers per bus), or there are so many buses it just makes traffic worse.

The CTA addresses those problems. They finish the section by saying they are “examining new and better ways to monitor and manage bus service, which we hope will reduce avoidable bus bunching, but CTA buses are as hostage to street traffic accidents and problems as much as individual motorists.” They seem to use traffic accidents as an excuse, but this isn’t that much of a phenomenon on the most-bunched routes, so it doesn’t explain all of the bunching.

To reduce bunching, there are a few remedies that would actually make a difference:

  1. Bus rapid transit: Proposed for Western and Ashland Avenues in Chicago. I do not consider the Jeffrey Jump to be “bus rapid transit” because it does not have dedicated lanes and signal prioritization for the majority of the route. Dedicated lanes that permit only public transportation vehicles are the only way to eliminate the problem of traffic accidents and congestion from a bus route.
  2. Signal prioritization: Part of any decent bus rapid transit route is signal prioritization, which extends a green light for an approaching bus or shortens a green signal on the cross street. The LA Metro Rapid bus routes use a form of signal prioritization (and Metro Rapid buses are a pretty quick way to get around Los Angeles). This works best when there is a dedicated lane for buses (aka bus rapid transit).
  3. Congestion charging: Politically difficult and costly to implement, but it certainly would get rid of drivers who wouldn’t want to pay! When all streets cost the same to drive on ($0.00), there’s no incentive for drivers to stay off of them. Equalize traffic by charging higher prices on more congested roads to achieve a traffic flow that makes it easy for everyone to get around and keeps the warrantless trips out.
Simple diagram of how signal priority works.

Simple diagram of how signal priority works. The same technology exists for emergency vehicles. Credit: Virginia Tech.

I’ve read some of the research on bus bunching (there’s quite a bit) and even the official response from the CTA seems to be using GPS and modifying the schedules to reduce bunching. But that doesn’t eliminate the problem, it just changes how it’s measured (which is just a way to make everything seem better than it actually is).

Paris bus lane

Bus/bike lanes: Paris has them. Buses in Paris are still slow, but that’s because of the street layout. Luckily, there’s a phenomenal subway system just a few meters below. Credit: Colleen’s Paris Blog.

None of the reading I’ve done makes an effort to point out how riders feel better about using the bus under these conditions. The CTA can explain why buses bunch and why drivers have to sit at green lights to regulate the schedule, but that doesn’t mean riders will feel better about it. The bus still is one of the lowest forms of public transportation partially because its mingling with private vehicles mucks up its reliability. Buses can carry many more people than the 3 private automobiles that drive next to them, and they should be prioritized forms of transportation – especially on the routes that have the worst problems.

Giving buses their own lane is not always feasible. Clark St, for example, is narrow enough as it is (which probably explains some of the bunching). It would be hard to add a lane just for bus service. But what if we could figure out where the buses are bunching along the route? Is there room before or after the bunch points where a lane could be added? The best way to find out what is causing the bunching would be to obtain data on where the bunching starts to happen to find a pattern.

Bunching on 8 of the 10 most bunched routes got worse from 2011 to 2012. The CTA has only been dealing with the problem when it happens, not trying to prevent it. GPS tracking may help to find the problem sooner and alleviate it faster, but it still happens all too often. A supervisor often has to meet the bunched buses to correct the situation. Measures like signal prioritization and bus lanes are expensive, but so are labor and pensions.

It’s unfortunate that more can’t be done without taking up an entire extra travel lane on some streets. It’s also unfortunate that we allow private vehicles to choke up entire bus routes that carry many times more people. That one driver trying to turn left to find parking could be holding up 60 passengers on a bus waiting to go forward. At multiple points per route, a few people have the ability to screw up a route, with effects that ripple across the entire network. That’s not transportation equity. If we want to make our cities better places to get around, there needs to be better investment in technologies that enable public transportation agencies to provide a better level of service as often as possible.

In your experience, what are some of the worst routes/locations for bus bunching?