As WalkScore's fine transit travel time tool languishes in extended beta, the alternative, Mapnificent, is getting some mainstream blog attention.
I may still be alone in this, but I as I explored with WalkScore's Matt Lerner here, I believe this tool, whoever finally perfects it, has revolutionary potential. It can easily be converted into a two-digit transit score which, unlike the WalkScore Transit Score, actually describes people's ability to get places. But it's bigger potential is as an alternatives analysis tool. When you city is facing a series of possible alternative transit projects, what if every citizen could use a tool like Mapnificent to see the exactly impact of each alternative on their mobility, and that of people and destinations they care about.
A major problem in transit politics today is that negative impacts of a project are obvious but benefits are often described in terms of ridership and development outcomes — things that don't matter to the selfish present-minded citizen. We will always have selfish present-minded citizens, and I'd rather work with them than complain about them. Until we help people see the way a proposed project will change their lives for the better, sensible transit projects will continue losing these debates.
It’s worth a read of this blog post to see how it works:
Specifically of interest is that it does use service headways, at least at a high level. It assumes you will time a trip to avoid a wait for the first service, and that subsequent transfers have a time of half of the headway at that time of day – which the author admits may or may not be accurate.
This is a great map tool.
1) I would like to see the default start time changed from 6 AM weekday to 12 noon Saturday. The reason: during rush hour, people that have to work are not really concerned about general mobility – as long as the one specific trip of getting to work and back is reasonable, nothing else matters.
On weekends, however, when people aren’t working, they have more time to go more places, which might be all over the city.
Choosing a default time of 6 AM weekday makes the transit system look better than it really is by giving them credit for service that only operates during rush hour, which doesn’t really improve mobility.
2) I really appreciate the bike-plus-bus option. I tried it out for my home as a starting point and adding the bike made a huge difference, which fits my real-life experience.
3) Assuming you can time the trip to not wait for the first service is optimistic. If the bus is late, you will wait, even if you do time it. There should be an assumption of a minimum 5 minute wait each time you board a transit vehicle.
4) I would like a feature to see a comparison between the mobility chart between transit (or bike+transit) and driving. The two charts could be combined with a shaded map based on the time ratio between transit vs. driving. For example, green => 0-1:1, yellow=>1:1-2:1, red=>2:1->3:1, Black=>3:1+. This illustrates which trips transit is doing effectively and which trips whey have problems with. This can allow you easily pinpoint bottlenecks in a transit system similar to a poorly-timed traffic light causing bottlenecks in a road system.
For example, one glaring problem with Seattle’s system that the map illustrates is a huge delay in traveling from anywhere north of downtown to anywhere south of downtown.
Actually, he uses headway divided by 3. I know that this would make my suburban transit agency look awful because it would add a 20 minute transfer penalty even though every connection is synchronized. WalkScore’s tool does a better job at this.
As for the 6AM thing, this is also unacceptable. Ideally, we’d want to be able to choose the time of day and week, but if not, then a weighted average could be appropriate.
Overall, the assumptions made are a pretty good start.
My general experience is that a timed connection is not reliable unless you are scheduled to arrive at the connection point a minimum of 10-15 minutes before the next bus leaves. In a world of 30 minute headways, this is barely better than the wait-time-equals-headway-divided-by-2 approach.
The problem is there’s just too many random things that can delay a bus, even when traffic on the roads is minimal. For example:
– A large group of riders boarding at once with everyone paying in cash
– A wheelchair getting on or off
– People asking the driver questions, such as which bus to take to get to X or what the fare is to get to X or what the transfer policies are
– Waiting for another bus to leave a stop before entering that same stop
– People loading a bike who don’t know how to use the bike rack
Conclusion: Unless the headways are really long (1 hour or more), headway/2 is good enough.
The assumption that you’ll plan your trip around the schedule of the first bus really discounts the value of frequent service… also, drawing circles for walksheds around bus stops is pretty close in some areas and way off in others.
On the other hand, the maps are pretty! The problems are pretty subtle, to me.
I guess the other thing with this tool is that it doesn’t include all the relevant agencies. In Chicago it’s missing Metra, which makes some south-side neighborhoods look utterly isolated, which isn’t true (although Metra and CTA are incompatible for transfers and Metra has mostly commuter-oriented schedules that are otherwise pretty sparse). It’s also missing Pace, whose inclusion would make graphs near far-out L stations look more full.
Missing agencies is one of the things that tends to affect non-local tools. If I was building a transit tool for Chicago I’d know my system wasn’t complete without Pace, Metra, NICTD, and the various water taxis (in something like that order, depending on the sort of tool — I hope I haven’t missed any!). Maybe even Amtrak for some sorts of tools (Mapnificent would have a little dot in Kankakee if you dragged the slider far enough to the right, which would be pretty cool). If I was building a national tool it would be a lot more important to cover a lot of different cities. Get the CTA schedule in, then move on to LA.
I like this tool. Good comments above and I really like the idea of comparing drive time… Measuring and understanding performance is critical to improving it. Visualization tech seems to be taking off in some really interesting directions.
Most projects take too long to complete for them to be “present” enough for “present-minded citizens”. People can and do move. I’m excited about Seattle’s light-rail extention to the University of Washington, but I also hope to be out of college and living somewhere else by the time it opens. (I don’t go to UW, but I do live near it.)
The concept is good – If you can use any combination of transit-walking to get where you need, you can get rid of that car.
The transit information (at least for my area) needs major updating.
Also , mapping for comparative purposes needs to be based on something else than zip codes . Given the mess that urban development was in US, my mailing address locates me in the “city” whilst my physical address is in a suburban area. For a suburban area I know I should have a better transit score compared with other suburbs , but it will definitely poorer if the scored compared with the ‘city’ …
Jared, I think your point about this being a great tool for helping average folks comprehend the impact of improving transit is a great one. I think that a lot of those of us (myself included, definitely) interested in transportation and urban planning take for granted our own understanding of space and time in an urban environment. While I may have imagined something like Mapnificent in my head on my own when looking at transit maps, I think it could be eye-opening for many people.
I have recently become very interested in the in the effect of information design and how information is presented for getting across complex ideas about transportation and urban issues. I think that Mapnificent is a great example of a simple, elegant tool for presenting and getting across a large amount of useful and interesting information.
This might be blindingly obvious to people, but one cool Mapnificent feature: it can automatically intersect its “clouds” around two or more points. Why is that cool? I can put a point where I work, a point where my wife works, adjust the sliders so they reflect our commute tolerances, then use intersect mode to get a map of places to look for apartments. It came up with some interesting ideas.
Unfortunately I had to fudge my work location, since I work in Snohomish County, and neither Google Maps nor Mapnificent has their transit agency on file.