Flutterby™! : Unsuck GGT

Next unread comment / Catchup all unread comments User Account Info | Logout | XML/Pilot/etc versions | Long version (with comments) | Weblog archives | Site Map | | Browse Topics

Unsuck GGT

2012-04-05 00:08:49.833355+00 by Dan Lyke 7 comments

After a Twitter exchange, David Edmondson of The Greater Marin Blog, had a challenge for me:

@danlyke Care to draft an Unsuck GGT piece?

I don't know that I have the background in urban planning to unsuck the entire system, so let's talk about what would unsuck it for me. I've put my proposal in the comment after the break.

[ related topics: Weblogs Bay Area ]

comments in descending chronological order (reverse):

#Comment Re: made: 2012-04-06 14:18:53.361481+00 by: Dan Lyke [edit history]

The thing I keep threatening to build a simulation to try to figure out is: Is there any circumstance with current development patterns where buses make sense?

The problem is that buses have a worse user experience for transport than cars by every measure, except (subsidized[1]) cost, and that's not enough to matter to most users. The only place they definitively win is in aggregate behaviors, parking and traffice, and that means you have to be traveling somewhere where parking and traffic is a problem. These are both temporally and geographically heavily bounded.

In order to make buses a better experience, you have to take resources away from cars:

And that bus-only lane would need to be physically separate (so you don't have the differential speed problems that you run into when your carpool lanes don't have enough density). This means building super-expensive overpasses to get pedestrians on and off the median of the freeway, which means... well... you should probably just give the extra lane to the cars.

[1] "Subsidized" called out up there because in practice in overall direct costs, buses are more expensive than automobiles per passenger mile.

#Comment Re: made: 2012-04-05 23:34:26.857196+00 by: ebradway

Don't even get me started on light rail. RTD just threw in the towel and is considering building special bus lanes along US-36 and CO-119. Based on the current (very, very overrun) budget, FastTracks from Denver to Boulder is pushed back to 2020 and Longmont to 2040.

#Comment Re: made: 2012-04-05 16:44:23.570618+00 by: Dan Lyke

Yeah, the problem is that any improvement only occurs if people are willing to take a near-term hit to their quality of life, and when that general improvement occurs we're back to the automobile being preferable to the bus once again.

The advantage that rail has is that the stops are easier to identify, the ride is more comfortable and therefore more usable, and, of course, the separate infrastructure. But even there in practice heavy rail has worse passenger miles per gallon rates than you'd expect, Sonoma County's SMART is talking about infrastructure capital costs that are 4x per passenger mile than the current 101 widening, and you have more payload for bikes. Hopefully.

On the other hand the SMART folks are basing all of their numbers on standing in cars, so there all of the advantages evaporate.

#Comment Re: made: 2012-04-05 15:48:48.991298+00 by: TheGreaterMarin

Not bad recommendations. GTFS is on the way - they had problems with inputing the every-other-day school routes - and realtime arrival data is also on the way, though likely not for another year or more. MTC pressure has a lot to do with that.

Stop location data must be included with the GTFS feed, though having the actual routing shapefile (like Muni) would help a lot, too.

@ebradway The problem of service reductions you're describing is likely due to states having just that attitude, that transit is a service for the poor who have no option. Rather than invest $5 million to save riders 15 minutes, a state is more likely to spend $5 billion on a new interchange to save drivers 2 minutes. Wisconsin has exactly this problem (thanks, Governor Scott!), as does Detroit. The irony is that faster transit service is cheaper transit service, as you don't need as many drivers or buses to maintain headways.

#Comment Re: made: 2012-04-05 14:22:00.979323+00 by: ebradway

Unfortunately, the driver's unions prevent many of these things from happening. The RTD (Denver Area Regional Transpotation Division) used to have an RSS feed with every bus's location updated every minute or so. That was in 2006. It was scrapped in 2007.

The only way to know where your bus is is to read the schedule, hope your watch is right and the driver wasn't too fast. And schedules are always variations on some sort of paper chart. No app that uses my GPS and says "there should be a bus coming in 15 minutes at this corner down the street".

I learned the challenge of providing rider stop information taking the trains in England. I understood the engineer perfectly but I had no idea where any of the placenames he was calling out were relative to where I needed to go.

My favorite move was two years ago when gas was holding steady at $4.50/gallon. The buses got really crowded. So what does RTD do? They announce service REDUCTIONS!

I've come to the conclusion that public transportation systems are designed for customers who have no other choice - a captive audience, per se. So the system can suck as much as it wants until their is public outcry. But the riders have little or no money - that's why they have no other choice.

#Comment Re: made: 2012-04-05 03:24:55.254773+00 by: Dan Lyke

Oh, I think it's also worth noting that in a decade, we'll have the first self-driving cars hitting the market. In two decades, those cars will be prevalent enough that we'll see about 3x the throughput per lane as we're seeing now.

#Comment Re: made: 2012-04-05 00:08:59.777516+00 by: Dan Lyke [edit history]

I used to live in Fairfax and commute into San Francisco. I have, recently, used Golden Gate Transit (GGT) to get from Petaluma to College of Marin to meet up with Charlene, so that we could drive home in the same car (Charlene commutes by car from Petaluma to COM). I have also driven around bus routes in Marin with a camera and GPS and note pad to help Charlene build better resources for teaching the developmentally disabled adults from 18 to 22 in the transitional program she works in navigate the bus system.

It's also important to note that GGT runs several local routes under contract to some more local transit agencies. I'm not sure what all the details are, when I refer to GGT I mean both the commuter routes in to San Francisco, and the local routes around Marin, although I'll try to keep distinctions between commuter and local, because they do have some different needs.

I do not know if "unsucking GGT" is possible. I have noted before that buses in 1999 cost per passenger mile what cars in stop-and-go traffic do in 2009, and though I can't find it on Flutterby, I know I've mentioned in various places that buses in the U.S. get gasoline equivalent passenger miles per gallon that's worse than single-passennger automobiles (you can start with the BTS data, remember to correct carbon emissions for diesel and LNG vs gasoline).

So, taking my use-cases:

Commuting
I used to take the bus to the ferry into San Francisco. I actually have no complaints on this front, except to note that when I had to fall back to the bus for the whole route, it was a genuine fall-back. People like to claim that transit miles are usable miles when you don't have to concentrate on driving, but that worked only if I got a coach bus with a smooth driver. If I knew that both routes were going to be on a bus, I'd consider driving and paying for parking. Even when the bus worked, the lack of good indications within the bus meant that I had to be watching out the window enough to be very aware of my location.
Recommendations:
  • In-vehicle "next stop" information. This (and the sign on the outside, you'd be amazed at how often that's wrong) needs to be automatically derived from GPS data, not reliant on driver intervention. You also can't rely on bus drivers announcing stops, often they're unintelligable (hilarious aside: Riding Atlanta's MARTA shortly after the Olympics, they had these brand new wonderful pre-recorded stop announcements, which would have been great had the train operator not cut in in the middle of them to unintelligibly mumble/drawl some confusing interpretation of the stop).
  • Buses in good repair, with some reasonable expectation that multiple buses would be run along high density routes at high density times (ie: no standing)
  • Accelerometer loggers and some sort of driver bonus for smoother driving
One-off trips
As I mentioned, I grabbed the bus from Petaluma to San Rafael, and transferred to the bus to College of Marin. I used 511.org to plan the trip. I didn't look at the times well and spent 45 minutes shivering in the cold at the San Rafael Transit Center. The driver didn't change the bus sign, so I had to ask the driver if indeed this was the right bus. I got off about two stops early, because I wasn't sure where the right stop was, but that's okay, because, walking, I managed to beat the bus to the next two stops and thence to my destination.
Recommendations:
  • Publish a General Transit Feed version of the schedule so that you aren't beholden to the bugs and glitches of 511.org. Really, how hard is that?
  • Fix your transit maps to show stops, so that we don't have questions about how many stops there are.
  • Reduce the density of stops, so that the constant loading and unloading of passengers doesn't destroy bus efficency.
  • Approach schedule and system map design from the point of view of the occasional user going to neighborhoods they aren't familiar with. Stop using PDF and designing for print! We're all using computers and smart phones, and PDF sucks.
  • Reiterating that previous "Tell us where the hell we are" request.
Building a user's guide for developmentally disabled users
This has most of the same components as building a user's guide for occasional/first-time users. And this is important because unless you get people comfortable with using the transit system, you won't. Charlene and I drove the bus routes that her students use, noting where each stop was, figuring the exact route where it differed from the GGT supplied maps, and taking pictures of landmarks and bus stops.
Recommendations: All of the above, plus:
  • Provide map and schedule in a KML file that can be overlayed on online map applications so that we can go see exactly where the bus stop is and what it looks like.
  • Get your freakin' stop location right. The transit map not only misses many physical stops, it puts a lot of them in the wrong place. This is makes planning difficult for people with limited mobility.
  • And, did I mention: Stop using PDF and designing for print! We're all using computers and smart phones, and PDF sucks.

Buses have a few potential improvements over automobiles.

Energy/Fuel efficiency:
In theory, buses can be very efficient. Greyhound claims some amazing numbers (although, as always, it's unclear if those amazing numbers are passenger miles per gallon or seat miles per gallon, many times operators will claim the former, using the latter numbers). In practice, efficiency will really only be found on commuter routes at commuter times, and even then much of the improvement will be the larger issue of reducing other traffic. A drum I've beaten many times is that our current policy of heavily subsidizing buses means that we're encouraging bad/expensive development patterns and wasting tremendous amounts of fuel. It is frankly unlikely that Marin will have the sorts of development patterns that make non-commuter bus service efficient before e-commerce and telecommuting change our transportation patterns such that we simply don't need to travel that much any more.
Income redistribution
The old saw is that it'd be more cost-effective and fuel-efficent to buy poor people cars than to provide bus service. At a micro-economic level, it's hard to poke holes in this argument. I'm not an expert in network theory, but I know that the people I know who live in low-income housing can't use transit because it doesn't stop anywhere near where they live. One of the things we need to make sure of if we're going to have bus systems fulfill this role is that buses are serving apartment complexes, and, probably more importantly, that new low-income housing is built near freeways and other transit thoroughfares. Right now, though, even with the heavy subsidies you generally have to be making less than ten bucks an hour to make the price difference between the $.50-60 that it costs to operate a car and the cost of a bus fare worthwhile.
Usable Time
The answer to the price difference is to make the transit time usable. A lot of transit system time is spent waiting for connections. I realize that transit center design is a compromise between making the place usable, and making it un-inviting enough that the homeless don't make camp and trash it. Maybe a compromise, as kind-of happens at the ferry terminal, is to have a room with tables, WiFi and sockets for laptops, that's accessible only to those holding a transit ticket. On buses, they have to have enough room to comfortably read or use a laptop. Standing has to not be a regular part of a route (it is, and, yes, I realize that this flies in the face of efficency), and they have to be smooth enough to make work doable (possibly accelerometer loggers and driver bonuses?). Otherwise even if the bus is as fast as driving, I'd rather drive.

Basically, these boil down to: Provide correct information. Provide it in a usable format that people can re-use and re-purpose. Focus on acquiring new users, because they're the ones you don't have.

Unfortunately, as long as the users of transit aren't the customers of transit (ie: most of transit is paid for by tax dollars extracted by special interest groups, not by the people who actually ride it), none of these improvements will be made. Insist that gasoline and property tax dollars actually cover the costs of infrastructure and vehicle operation, and the market will solve the grander problems so fast your head will spin.

But at the very least, with just a little information redesign, GGT could suck a whole hell of a lot less. And with just a little more hard infrastructure (ie: waiting space at the transit centers), or even with some public-private partnerships (a coffee shop right next to the transit centers), I wouldn't say "I'm never taking the bus from Petaluma to COM again".