chicago: a draft frequent network map

My case for Frequent network mapping continues to inspire tech-savvy readers to draw their own.   Jeff Wegerson, for example, has developed a method of automatic map generation using the same data files that Google Transit relies on.  He's tried it out on Chicago.  His entire map is here, including technical notes for others who'd like to try it.

The most interesting thing to me from Jeff's work is this nice illustration of the threshold definition problem.  Exactly how frequent is frequent?   If you say every 7.5 minutes, Chicago has exactly one frequent service, a piece of busline on 79th Street.  If you say 15 minutes, you get a network covering most of Chicago.

If you stare at these, there's a lot of information here about how Chicago's transit system really works, information you'd never get from a standard system map or from reading timetables.  Who knew that a portion of 79th Street is more frequent than any of the rapid transit lines?  Who knew that east-west services on the south side are more frequent than the north-south services there, so that when you draw the limit at 11 or 12 minutes you get a bunch of parallel lines that connect with rapid transit but not with anything else?  You can also see easily that the territory of the Orange Line corridor (the string of green dots extending southwest of downtown toward Midway Airport) is a relative dead zone by Chicago standards, supporting north-south 15-minute frequencies but not much better than that, and no east-west lines to speak of.

In fact, as someone who doesn't know Chicago all that well, I'll remember these maps next time I need to visualize the city, becuase they tell me what really matters about the city's patterns of transit demand, which in turn tells me a lot about the city's shape.

As for Jeff's graphical choices:  How do you feel about rendering rapid transit as a string of station dots?  I like it, as long as it's clear how the dots are connected.  In a very dense rapid transit system like that of Paris, it would be chaos, but in Chicago's simple radial rapid transit network is not hard to follow.

Jeff also uses a convention that I've seen in other grid cities but never quite understood:  different colors for north-south lines as opposed to east-west lines.  Los Angeles used to do this on its maps too.  I've always assumed that the purpose of color (and line-weight) in mapping is to convey information that is not obvious from the lines themselves -- information such as frequency and span.  But I can see that a north-south line is north-south, so why do I need a color to tell me that?

There's clearly some viral potential in the frequent mapping idea.  I've heard from readers who are working on automatically-generated frequent network maps as a business venture.  I also had a brief exchange with Google Transit's staff on the subject, and got their standard "we only have so much time" response.  So I hope someone makes a business out of this, because if they succeed, Google will probably buy them out, and they'll be set.  If that turns out to be you, remember the humble blogger who lit the spark!

Comments

[ant6n]

I guess there is this general question whether frequent network maps should be automatically generated, or made by hand. Which is probably related to the question whether they should be abstract and compact, or geographically accurate. Or whether they should be published today, or ... later. It took me more than a solid week to make a map of Montreal's network by hand.

But I feel that there is a certain local culture associated with a city's transit map. The map and the city share some identity of sorts - just look at London's tube map. So from that point of view having a person create that special map might be a good idea.

I do agree though, that google maps could really show a frequent transit layer on top of their map...

[Nwali]

I am also trying to draw the City of Tshwane bus routes on google maps. Its my first time doing this and I am encouraged by Jarret. Thanks Jarret. For those who want to see the map can visit this link:

http://maps.google.co.za/maps/ms?ie=UTF8&hl=en&msa=0&msid=101582471106944558287.00048f2f5114e5c1aafe4&ll=-25.744458,28.1908&spn=0.002358,0.003449&t=rm&z=18&iwloc=00048f43ab6dfaa61aa91

[Joe Clark]

And again, the wrong possible colouring: Nobody with either the weak or strong forms of colourblindness (deuteranopia and protanopia) will be able to tell the lines apart.

Sure is purdy and works great for the creator, so problem solved, though, right?

[Tom West]

On colours... I agree with joe clark that the choice of colours is poor for people who are colourblind, but that's easily fixable. (cyan/blue/purple/magenta/red is a set of colorus which work well).

Jarett makes a good point that colouring lines by direction seems pointless, because the shape of the line gives you its direction. However, this map isn't plotting liens - it's plotting stops. Therefore, the colour is needed to tell you the direction of the line through the stop, because the stop isn't shown.

[Zoltán]

In the discussions on frequent service mapping, I feel that there is at times an either-or debate between diagrams showing only the frequent network, and including the frequent network on ordinary network maps. To my mind, any city ought to do both. This isn't so unreasonable; any network map clearly could and should highlight frequent services, and any city could and should have a simple map of frequent services.

Leeds does this, though on the initiative of the private operator that has more or less a monopoly on city services. Its network map shows bus lines in grey but shows frequent services with an individual line colours, and also has a diagram only showing those coloured lines. (See the first two maps on http://www.firstgroup.com/ukbus/yorkhumber/leeds/map/index.php ). This would be a perfect pair of navigational tools if only it didn't miss out the one frequent corridor provided by another private operator, a problem not facing most American and European cities.

[Matthew]

I've lived in Chicago for about half my life, and I've never been as far south as 79th, so the frequency of that route comes as a total shock to me. It's also interesting to me how quickly the frequency peters out going south from the Loop. There's actually quite a bit there, with much of Pilsen falling into the dead zone on the northern side of the Orange Line corridor (along the Pink Line), and that might say something about the relative amount of clout that the newer Hispanic populations have in city workings (Pilsen is the city's largest predominantly Mexican neighborhood).

Tom West makes a great point about stop color, since about every fifth stop location is essentially shared both north-south and east-west, given the structure of Chicago's buslines.

[Daniel Howard]

Yeah, I was wondering why the L lines were a series of circles when I realized everything was a series of circles, and that yeah, this isn't line service frequency so much as stop service frequency. But yeah, that 79 bus has a lot of service frequency, looks like. I think that is where they once had branches off the North-South line service the stock yard and Pullman. Or, maybe this could be an artifact of maybe they route buses to a barn down this street?

[Alon Levy]

It's strange that 79th is the most frequent bus. Doesn't Ashland have the highest ridership? Or is it dispersed along more route-km than 79th?

[Jeff Wegerson]

COLORS. Yes I became aware from this blog to watch my colors. Actually I became aware from a different exercise where I created a table of information for another subject based at first on colors. Then a reader with ever more sever eyesight limitations took me to task and I changed the work to show information by both colors and font styles.

I chose the colors on my own and without reference to anything else because I wanted to show slower parts of the system as a slower color. But because sometimes lines shift to other parallel streets I wanted to have some indication of that shift. Therefore the similar but different colors for different directions.

The map above could easily be done in a single color with different shapes designating different kinds of service.

AUTOMATIC Generation vs hand generation. Actually there is probably a weeks worth of work or more in my "Automatic" map. I don't think we are dealing with an either/or situation here. This is an exercise not a finished product. A finished product will likely always require a significant amount of hand work.

79TH STREET was a surprise to me as well. My take is that this is where the more intense and older poorer parts of the far south side have given way to more middle class parts. To me it seems to say we need BRT on that corridor.

CITY OF TSHWANE - Instead of using Google Mapping you may want to try a Structured Vector Graphics piece of software like the Open Source "Inkscape." I have more technical details at the link provided and would be happy to fill in any technical holes left by my brief synopsis.

[Aaron Brown]

If you look at the transit stats in Chicago, the 79th Street bus actually has the highest ridership in the entire system. I think it's due to the fact that it's a long route through many fairly dense neighborhoods with lack of rapid transit options to compete. I'd guess that many of these rides are toward the Red Line or the multiple Metra lines it intersects.

To Jeff's point above, the route was actually one of the CTA's proposed BRT routes last year, but apparently fell behind the Jeffrey Express when it came to getting federal money.

I actually think the most interesting point is the lack of frequent transit on the Near / Mid South Side. Given that many of these neighborhoods have begun to (slowly) gentrify (think Pilsen or Bridgeport), it's surprising that they aren't better served by the CTA. Also seems like a missed opportunity to spur infill development.

[observer]

I'm not sure that similar maps could not be constructed by looking only at ridership numbers. This became apparent to me when I checked the official ridership statistics on the rapid transit lines and where these lines appear on the frequency map. Simply put, the heaviest-used lines get the most frequent service. Well, duh...

[You will find selected stats at the end of this post]

I also propose that the dead spots, by the Orange Line, for example, are no such thing. This line passes through a mostly industrial areas where few people live into a low-density area of single family houses. If you don't believe this is true, use Google Earth to trace the Orange Line as it exits the Loop elevated structure on its way to Midway Airport. This line which follows a tortuous path because it uses obsolete railroad tracks opened in 1993 and while a better path could have been devised to serve more people, I doubt that if could have been built because of the cost. There are only 8 stops after leaving the Loop which if you are on the Red Line heading north from the Loop would get you to Belmont.

There are two high-density neighborhoods on Lake Michigan on the south side, Hyde Park and South Shore. The 79th Street bus connects the CTA Red Line and two Metra electric lines to South Shore.

The South Loop has indeed been gentrified but most of it was built on unused railroad tracks and is mostly low density townhouses and single family houses. Those living north of Roosevelt Road can, if fit, walk to work and those south cab take a number of buses or the Red Line. Again, you can verify this by using Google Earth.

Enough speculation, here are the promised statistics:

Rapid Transit Daily Ridership (July 2010)

Red 240,000
Blue 150,000
Brown 94,000
Green 64,000
Orange 55,000
Pink 28,000
Yellow 5,200

Bus Ridership Over 20,000/Day (July 2010)

29,500 79 79th Street (e/w)
28,400 9 Ashland (n/s)
26,000 26 Western (n/s)
24,500 66 Chicago (e/w)
22,700 4 Cottage Grove (n/s) South Side Only
22,500 151 Sheridan Road (n/s) North Side Only
21,700 77 Belmont (e/w)
21,200 3 King (n/s) South Side Only
20,400 20 Madison (e/w)

[Matthew]

@Aaron Brown: I keep hearing that Pilsen's gentrifying, I even work there, but all I see are either people who work in Pilsen or else Hispanics. Beyond a handful of restaurants and the art galleries, though, I haven't seen much there that would be considered "gentrification."

[Matthew]

@Observer: A similar map in appearance could be constructed, but there would be no guarantee that it would be a frequent network map. Also, while the Orange Line runs through industrial areas, it's often within walking distance of under-served neighborhoods. One example is the Halsted Orange Line stop. It's only 10 minutes from a fairly populated area to the north, and even closer to the south, so while the line itself runs through industrial areas, it isn't far from residential areas that don't have proportional service. I think the dead zone isn't necessarily indicative of a lack of people, but rather a lack of public transit routes that allow a rider to get anywhere useful in a reasonable time frame.

[Aaron Brown]

@Matthew.

You're right that Pilsen is nowhere near to being fully gentrified, my point was that it's started to see a bit more restaurants, shops, new housing and middle/upper income residents. I live in Pilsen right now, and there are definitely a lot more places to eat / shop and a much wider mix of residents than there were 10 years ago.

The recession definitely slowed down the gentrification process, but I would bet that a neighborhood that close to the loop, with a solid retail strip (18th St.), and its own rapid transit line won't take long to become hot again. And to my point above, if transit options improved for East Pilsen (other than the Halsted bus, there's not much), I think the area would be even more attractive.

[Jeff Wegerson]

@Observer: My first version of the Chicago map used the number of trips as a proxy for frequency. Jarrett objected and insisted that I stay true to the idea of consistent service even through the midday. The objection, again, of using other measures for frequency, which applies to ridership as well, is that you can have bunched trips and riders around the rush hour and still have sparse frequency in the midday. And vice versa.

Indeed the yellow line at the bottom of the el lines at 5,200 riders actually is on the ten minute frequency map.

[Anon256]

An hour or two with the New York City Transit GTFS files, Perl, and GPSVisualizer.com produced this map of all bus stops in New York City that see an NYCT bus at least every 10 minutes between 6:30am and 8pm (i.e. comparable with daytime subway lines). Obviously this is not a format useful to users of the system, but being able to quickly and automatically generate a view like this might be a useful starting place for drawing a frequent network map.

Focusing on stops, rather than lines, does lead to a few aberrations, such as major transfer points in Staten Island showing up when there isn't actually frequent service from that point to anywhere in particular, and the northern portion of First Ave in Manhattan not being marked as a few M15 trips slip to 11 minute headways by the time they get up there. On the other hand, working based on stops makes it much easier to see instances like Fifth Ave where several routes together provide frequent service.