maps and aesthetics: washington’s hidden spiral

Transit maps always express a choice about how you see the city.  Do you want to show the city in its geographical detail?  Or do you want to be able to show the structure of the transit system, which involves expanding some areas and reducing others, often leading to distortions of scale that mislead the geographically-minded rider?  Like many, the classic Washington DC Metro map does this, shrinking outer distances and exploding inner ones.


Structure can be rendered many ways, and once you're free of literal geographic scale, it's tempting to create some other visual logic.  Do you want to emphasise the concentric quality of your city, or do you want it to display many equally important points?  Which is bigger, the lines or the stations?  Do lines meekly serve stations, or are stations mere decorations on lines? 

Even more basic, what kind of structure makes you happy?  The designers of the Wellington, New Zealand transit map like diagonals, rounding all routings off to the nearest 45 degree angle.

  Wlg slice

It sacrifices certain geographical information to show the system in a certain pleasing way, which is fine. 

Point is, you can find any balance of geographical accuracy, systemic clarity, and sheer visual pleasure, and still be accurate.  As for whether it's useful, that depends on the audience and purposes.

So there's nothing technically wrong with mapping Washington DC's metro system like this (follow link for sharper one):

Bossi spiral

… as Andrew Bossi does.  As a system map, it's a strong visual choice, but it's not inaccurate!

10 Responses to maps and aesthetics: washington’s hidden spiral

  1. Brent Palmer July 19, 2011 at 8:48 pm #

    Obviously, the spiral map is all about narrative navigation, whilst most schematic “maps” are a mixture of narrative and spatial.

  2. GD July 20, 2011 at 1:03 am #

    as much as the DC map distorts scale, it still gives you some idea of where you are in the city – I’ve once navigated a car through Paris’s outskirts using a similar mini map. I find those maps useful, at least once I have an idea of what the physical shape of the city is.
    IIRC, there seemed to be agreement in an earlier comment thread that every metro station should feature a standard-projection full system map. If the Metro map does at least retain some semblance to the physical layout of the system (as D.C’s does) then you have a lot of an easier time getting oriented on the larger, full-system thing.

  3. EngineerScotty July 20, 2011 at 7:30 am #

    Oh noes! The transit death spiral!

  4. Alexander Craghead July 20, 2011 at 11:24 am #

    The thing to remember is that the purpose of tranist maps — and in fact most maps — is data visualization, not representation. Accuracy to geography, topography, and such is (usually) not important or sometimes even relevant at all. While the spiral is funny, it fails because it fails to convey an important piece of data, the relationships between the various routes.

  5. Jarrett at July 20, 2011 at 1:19 pm #

    Alexander.  On the contrary!  It very accurately portrays all the connections among the routes!  It's just not interested in the relationship to the city!  For what it its, it's perfectly accurate.

  6. Wai Yip Tung July 20, 2011 at 3:05 pm #

    This is a great demonstration of Topology. Mathematicians will be delighted 🙂

  7. Andre L. July 20, 2011 at 4:48 pm #

    One aspect concerning structurally-eased maps, when they depart often from geographical positioning of stations and routes, is that they might mislead passengers making connections (transfers).
    Indeed, suppose someone is travelling on Washington from Ford Totlen to Rosslyn. Minimizing the number of transfers, there are 3 options:
    I. Red Line to Metro Center, change for Blue or Orange Line to Rosslyn
    II. Green or Yellow Line to L’Entant Plaza, change for Blue or Orange Lien to Rosslyn
    III. Yellow Line to Pentagon, change for Blue Line to Rosslyn.
    How should a person without deep knowledge of the system (like frequency of each line, time spent walking on transfer stations) could address such a scenario (which is not uncommon at all in cities with a interconnected, intermingled network of frequent rail service)?
    I suppose most people will try to find the “most direct and straightforward route”, which would be option I. But the schematized map might betray an user if they bear not much relation to real geographic placements.
    Other alternative would be counting the number of stations: option I involves travelling through 10 stations, option II – 14 stations, option III – 10 stations.
    In a system like Paris Metro within the most central districts, using the number of stations passed as a proxy for travel time might work, as they are more-or-less evenly (and shortly) spaced, putting the burden of time travel on stopping and embarking/alighting. However, in a place like Washington, that might be not the case at all.
    I know the example I gave is somehow overstating a problem, as a frequent traveler would likely learn, empirically, what is the best way to travel between those two stations rapidly (maybe option II has the shortest walking distance on transfer point of the less crowded trains…). Yet, distorted maps might create more serious trouble when one is facing trips on the “compressed” zone, or on schematized bus route maps.

  8. Alon Levy July 20, 2011 at 5:14 pm #

    The main problem with the spiral map is that it doesn’t actually show topology. It shows the line-station relations, but not the way the lines intersect. A topologically accurate map should be topologically equivalent not just to the graph of the geographic map, but also to its embedding in two-dimensional space. Put another way, such a map should distinguish between the cruciform configuration in which both lines are straight, and that in which both lines turn 90 degrees at the intersection.

  9. ant6n July 21, 2011 at 7:14 am #

    A geographic map won’t help you in many cases – it might even deceive you into thinking some path is shorter than another, because your map is “geographically accurate”. Consider the case of Manhattan, where the A train is much faster than the 2. Even if you take the local 8th ave train, the speed difference is quite noticeable, still.
    In many cases, schematic maps will only really distort suburban areas – where there’s only one line anyway. And then just using distances to compare routes might work out fairly well.
    On the other hand some schematic maps, through the distortion of areas and omission of some geographic features (like streets), have the space to provide information that is potentially more useful than knowing geographic distances accurately. These can include providing a better sense of the stations (where are the transfers long? Untangling where many lines meet), information on fare zones, etc.

  10. Blackurbanist July 27, 2011 at 9:22 am #

    Both are needed. Give the spiral to us transit heads that flip out over the distortion and need a more advanced experience. Keep the geographic map for tourists, but also put a real map next to it. Many Metro stations have the simplified Metro map next to the bus map that actually has a real scale with all the streets and landmarks. Also, hotels hand out more detailed maps that include both geo scale and stylized maps. Folks that still can’t read that are just completely lost.