Monday, August 31, 2009

Fairfax conducts economics experiment

As part of the current fiscal years’ budget balancing measures, Fairfax County raised the price of express bus routes 380, 595 and 597 from $3.00 each way to $7.00 each way.  Their exact reasoning was not publicly stated, though I believe it had something to do with the recent increase in federal transportation subsidy from $120 per month to $230 per month.  With the transportation subsidy, round-trip travel less than about $12 per day is essentially free, providing a means for local governments to get additional revenue without losing much ridership.  This depends on there not being close substitutes and on your ridership being primarily government workers.  On two of these express bus routes, Fairfax found this combination and was able to nearly double revenue.  The third one saw ridership plummet.

Because this provided an interesting economics experiment, I obtained ridership data from Fairfax County for the increased fare lines, as well as three other lines unaffected by the change.  I obtained data for the months of June and July.  Other than the change in fares, there was no change in service frequency or schedule.

The baseline data showed a slight increase in ridership from June to July.  Average ridership grew between 1.75% and 7.8% for lines 171, 401 and 950.  There was not a change in fare for these lines.

fairfax connector ridership study unaffected routes For lines 595 and 597, which travel between park and rides in Reston and Pentagon/Crystal City, respectively, the lines experienced modest drops in ridership, about 20%.  Combined with nearly doubling of fares, the change resulted in a net revenue gain for the county, nearly doubling.  This is probably due to the high government employees ridership and lack of close substitutes for these lines.  The nearest substitute other than driving would be to take a bus to West Falls Church, take the Orange line and then transfer to the Blue line, a “three seat” ride taking approximately the same amount of time (but with three vehicles, it’s more likely that one of them would have a problem).

In contrast, ridership on line 380, from Franconia/Springfield to the Pentagon, was cut by almost 80%, and average revenue was cut in half.  Although the ridership is likely a lot of government workers, the Blue line provides the same trip with a much better frequency, about the same travel time, a more comfortable ride, and after the fare change, about half the cost.

fairfax connector ridership study These results show that it’s sometimes possible to get more revenue by raising fares, but the results depend heavily on the circumstances.  Fairfax should reconsider its decision to raise fares so heavily, at least on the 380.  Where there are close substitutes, large fare increases can actually decrease revenues and are counterproductive.  Now there are fewer people riding the 380 and paying more each, so the county has to subsidize the bus line even more.

UPDATE:  The county has informed me that they are studying a new variation of the 380 line to visit more park and rides and charge a lower fare.  This is in the preliminary stages so it would still need to be vetted by Supervisors for impact on their constituents.  It is still a very, very rough draft plan.

Wednesday, August 26, 2009

Metro trip planner problems

Other than wanting to find out if they could be paid, the other reason Metro said they were not working on partnering with Google for their transit mapping service was the quality and availability of Metro’s own trip planner.

Most of the time, when I use the trip planner for rail trips, the results are fairly accurate.  The interface is clunky but improving, and it seems to assume an unreasonably long time to make a transfer.  This is probably a design decision to ensure that even the slowest-moving customers can catch their train.

I recently was invited out to dinner at a restaurant at Fairfax Corner, out in the Fair Oaks area (see map).  Using the Metro trip planner gave some interesting results, to say the least.  If you’d like to try this, the search was from “Eastern Market” to “4250 Fairfax Corner Ave [Fairfax, VA]” leaving at 4:20pm on a workday.  Here are the trip planner’s suggestions:


View Larger Map

1.  Take the Orange line to Vienna, wait 28 minutes to take the 623 past the restaurant (there’s a stop about 0.3 miles from the restaurant), then get off at a park and ride and wait 24 minutes for the 605 going back the way you came, stopping at the restaurant you passed.  Total time: 1:59.

2.  Take the Orange line to West Falls Church, wait a few minutes and take the 505 to Reston, then wait 49 minutes to transfer to the 605 to the restaurant.  Total time:  2:50.

3.  (This is the really funny one)  Take the Orange line to Vienna, wait 9 minutes to take the CUE-GOLD bus in a closed loop, arriving back at Vienna, then wait 57 minutes to take the 621 to the restaurant.  Total time: 3:03, of which almost two hours is pure waste.  It might as well have said, “Go to the Ugly Mug and drink for an hour and a half, then take the Orange line . . .”

I looked into how stable this trip was, by varying the start time.  Earlier departures resulted in earlier closed-loop rides on the CUE bus system, until the trip planner stopped providing that helpful suggestion (it only gave two options at that point).  Later trips shifted to suggesting that instead of continuing on the Orange line to Vienna, I should get off at East Falls Church and take a bus (the 2) to Vienna (which takes 50 minutes instead of 10).

The real solution, after studying Fairfax Connector’s bus map (large PDF), three bus schedules and calling the customer service phone number, is to take itinerary #1, get off a little beyond the restaurant and walk back along the route.  It’s less than a half mile and I’m in fair shape.  Takes about an hour and a half.

Of course, without Google Maps providing transit directions, it’s not a given that Google would do any better.  But Google has the incentive and resources to get their trip planner right.  They’re serving trip planner results for most of the country now, and problems with their search algorithm would affect many more riders than just Metro.

Others were able to point out very strange results from the planner, such as the suggestion to take a bus, Green line and Yellow line trains from “Greenbelt Center SC” to Huntington Metro station.  The routing was correct, but it took over 7 hours for the train to get from Greenbelt to Fort Totten, and over an hour to get from Fort Totten to Huntington (This was searched on 8/17/2009).

What’s the strangest thing you’ve seen?  Is the trip planner helpful?  Please share your funny (strange or ha ha) trip planner results in the comments.

Monday, August 17, 2009

Metro posts bus information online

Metrorail stations contain very informative bus maps showing the routes around a particular station. We recently looked at those at asked what additional bus map information you would find useful online. Metro decided to take our suggestion and post the station bus information maps online on each station's webpage. They haven't posted all of the information that you can get in stations, but there's a lot there.

For example, here's the map from Dupont Circle (PDF):


One of the most useful part of the map is the table of bus frequencies. It shows which buses just aren't operating late at night on the weekends. It also helps sort out the buses that travel once per hour (you should probably consult a schedule or trip planner and show up early) compared to the ones that travel every 10 minutes (just show up). This is great for people that are unfamiliar with the bus system but live or work near a Metro station. The station may have great service by bus as well as rail, but it's not always clear from the large system map what lines are available at what times, or which lines travel frequently enough to be convenient.

The Dupont map also shows, in yellow, the routes of buses that carry travelers from the station. So far, many other stations' maps don't contain the yellow lines, and instead just display a general downtown bus map. In the future, it would be great if Metro were able to post the same maps for other stations, which they already have in stations. But the other stations' maps do contain the tables, diagrams showing the locations of bus stops near the station, and more.

Metro's starting to develop the kinds of information technology that help people learn to use the bus system to its full potential. We have the full system maps to give the overall picture, the station maps to help people that frequent a station take a bus instead of driving to the station, the online trip planner to help plan trips, and NextBus to make waiting for the bus more productive and less unpredictable. Pretty soon, we'll have flash passes available on Smartrip, automatic reloading of fare media, and online Smartrip management, which will let riders put fare payment on autopilot if they so choose.

What other information should be online? I'm angling for a "12-minute" style map. If I had more time and better data-manipulation capability to deal with the 20 megabytes of text files in the Google Transit Specification data, I would be looking to make custom "stem and leaf" bus schedules.

Crossposted from Greater Greater Washington

Wednesday, August 12, 2009

Metro plans Bike/Pedestrian Study

Metro's planning department recently held a community meeting to seek input on bicycle and pedestrian access to Metrorail stations. If you're interested in providing your input, please take the online survey.

Over the next 20 years, Metro ridership is expected to grow significantly (up 40% from 2005). Kristen Haldeman, the project manager, said that the expected growth would make the Inauguration Day crowds a daily experience. It's going to be impossible to handle that kind of ridership if people try to get to the Metro stations with the same driving mode share we have today. There aren't enough parking spaces (parking is about 90% full on a typical weekday, with some VA stations at over 100% full*), there isn't much land to build more parking garages, and Metro doesn't have the money to build more garages.

That's why Metro is starting this project. They're concentrating on stations where a large portion of people driving come from very short distances, typically 3 miles or less, and figuring out what Metro or local jurisdictions can do to encourage them to walk or bike. Sometimes the solution will be fixing some sort of connectivity issue, like a nearby trail that doesn't have an obvious or safe route to the station. Sometimes it will be just a matter of providing signs from the station to the trail or vice versa.

They're also looking at improving bike storage at stations, with better lockers, bike cages or improved or expanded racks. They're also looking at coordinating with Smartbike at downtown destinations so you can ride from your final station to your destination. A lot of these improvements are going to go into Metro's 10-year capital needs inventory, and be part of the recommendations staff makes to the Board for capital funding.

The better portion of the evening was spent discussing the consultants' station typology. They took all stations in the system and tried to find similar characteristics and an example station. Then, they will develop recommendations for those example stations and see if those recommendations are applicable to the others in the category. The categories are given in the presentation. If this method is successful it will save Metro and the consultants a fair amount of time, so it's important that the categories and example stations be selected accurately.

A look at the categories (PDF) illustrates the wide variety in stations, from freeway-dominated Franconia-Springfield, to towering Ballston, from "campus" oriented Medical Center to suburban West Hyattsville. The consultants were especially interested in getting feedback about the categorization as well as the selection of example stations, so please take a look at that and provide it in the comments.

Metro is interested in your opinions on access to Metrorail stations by bike or foot, the survey is collecting data. I'm going to point out this article to the Metro project manager so if you leave comments here she might read them.

The project team will collect data until August, then investigate the model stations and best practices from other cities. They'll draft recommendations and develop cost estimates until November, then an implementation strategy in January. By March, Metro expects to have a final plan to present to the community and the Board.

*Meaning that some people leave during the day but others come and take their spaces.