Redesigning the New York City subway map

The long and complicated path that led to Eddie Jabbour's KickMap.

By Julie Steele
July 27, 2010

Note: The field of data visualization is much broader than most people conceive of it, and exploring this breadth was one of our primary goals in compiling the projects described in “Beautiful Visualization.” In the following excerpt, KickMap designer Eddie Jabbour explains the complexity he faced and the trade-offs he made while reinventing one of the most iconic maps in the world.

What follows is Eddie Jabbour’s story, as told to Julie Steele:

Learn faster. Dig deeper. See farther.

Join the O'Reilly online learning platform. Get a free trial today and find answers on the fly, or master something new and useful.

Learn more

Maps are one of the most basic data visualizations that we have; we’ve been making them for millennia. But we still haven’t perfected them as a tool for understanding complex systems — and with 26 lines and 468 stations across five boroughs, the New York City subway system certainly is complex. The KickMap is the result of my quest to design a more effective subway map, and ultimately to encourage increased ridership.

The need for a better tool

I was born in Queens and raised in Brooklyn. The first subway map I saw was my father’s, circa 1960. It made a vivid impression on me because it intimidated me. I saw a gray New York with red, green, and black lines running all over it like a grid (see Figure 5-1), and hundreds of station names attached (1). It reminded me of a complex electrical diagram that I couldn’t understand; it looked very “adult-serious” and even a little scary. I hoped I’d never have to deal with it.

Figure 5-1. The 1958 New York City Subway map designed by George Salomon. 1958 New York City Subway Map © MTA New York City transit. Used with permission.

London calling

In college I majored in design, and I spent half a year studying at the University of London. I was all on my own in a huge city I had never been to before. I quickly learned that the London Underground was the way to get around and that the “Tube map” was the key to understanding it. That map (which of course is the acclaimed Beck map seen in Figure 5-2) was brilliantly friendly: simple, bright, functionally colorful, designed to help users easily understand connections between lines, and physically tiny. Folded, it fit easily into my pocket, to be whipped out at a second’s notice for immediate reference (which I did often!).

Figure 5-2. Harry Beck’s map of the London Underground makes a complex system appear simple and elegant. 1933 London tube Map © TFL from the London transport Museum collection. Used with permission.

London was a medieval city, and therefore its street pattern is random. You cross a crooked intersection and the name of the street you’re on changes. There’s no numbered grid to provide a frame of reference (like in New York), and moving through the city can be a disorienting experience. The genius of the Beck map is that it makes order out of this random complexity, with the River Thames as the only visual (and geographic) point of reference to the aboveground world. And for that reason, the map’s layout is iconic: when you think of London, you probably think of that Tube map. But even as a design student, I didn’t think much about the form of it at the time — it was just so simple and easy to use that travel felt effortless.

The combination of that effective little map and my unlimited monthly “Go As You Please” pass allowed me to use the Underground daily to explore London. I went anywhere and everywhere with ease and got the most that I could out of that great city. The Tube map imparted information so quickly and clearly that it became an indispensable tool and an integral part of my experience. It made me feel that London was “mine” after only a couple of weeks of living there. What a fantastic and empowering feeling!

In fact, I formed such a warm attachment to that valuable tool that at the end of my stay, just before I left the city, I went to my local Underground station and got a brand new Tube map, and when I returned home to New York I had it framed.

New York blues

When you come back to your own city after six months away, you look at everything with new eyes. When I got back to New York, I saw our subway map — really saw it — for the first time since I was a kid. And I thought, compared to London’s, our subway map is poorly designed.

I remember thinking that the New York subway map was the opposite of the Beck map: huge in size, unruly in look, cluttered, and very nonintuitive. I realized that this map was in many ways a barrier to using our great subway system — the opposite of the Tube map, whose simplicity was a key to understanding and using the Underground.

Even as a designer, however, if I ever thought of creating my own subway map I must have quickly dismissed the idea. This was in the late 1970s, and I’m not a T-square kind of guy. The amount of discipline and mechanical time it would have required for anyone but an experienced draftsman to undertake such an endeavor was unthinkable in that precomputer era.

The map’s deficiencies left my mind as I pursued my design career. Like most New Yorkers, I used the subway map rarely and never carried it. This was in part because of its size: it was as large as a foldout road map. If I needed the map’s information to get to a new location, I would tear out the relevant six-inch square portion from a free map in the station and throw the rest of it in the trash! I often saw tourists struggling with the physical map and felt bad for them, remembering my great experience as a student in London.

Better tools allow for better tools

Now, fast-forward to one night years later when I was taking an out-of-town client to dinner at a downtown restaurant. As we waited for the train, he confided to me that New York’s subway intimidated him. I was surprised: the crime and grime of the 1970s-1990s were virtually gone from the system, and I was proud of our shiny new air-conditioned cars and clean stations. But in our conversation on the way downtown, I realized that his fear lay in not being able to decipher the complexity of the system: all the lines and connections. That’s when I realized that the problem for him, too, was the map. My client was very well traveled and urbane; if he found the system intimidating, then there really was something wrong with the communicator of that system — the map.

At that moment the subway map re-entered my consciousness, and it hasn’t left since. It was 2002. I had my own design agency and my own staff, each of us with our own computer loaded with a copy of the greatest and most elegant graphic design tool available. I realized that now, just one person using a graphic design program like Adobe Illustrator had the power to create his own subway map! And I challenged myself to do something about the map.

Size is only one factor

When I decided to try making a new map as a weekend project (ha!), the first thing I considered was the size. Since the New York City subway system has about twice as many stations as London’s, I decided to give myself twice as much space as the Tube map takes. (Even doubling the size of the Tube map, the result was about one-fifth the size of the existing New York subway map.)

First, I took a paper version of the official Metropolitan Transit Authority (MTA) map (a version of which is shown in Figure 5-3), cut it up with scissors, and put it back together in a more efficient way (literally with Scotch tape), just to see the possibilities. I was encouraged as I managed to reduce the area by more then half. Gone were the 56 bus pop-up boxes and other nonsubway information! Then came the laborious task of creating an actual map. I entered all the station names and lines into an Illustrator document, and in two months, voilá! I had my very own smaller map! I folded it and easily put it in my wallet, and I carried it around and showed it to all my friends. They liked its size, but of course nobody wanted to actually use it, because it still had many of the major design issues that made the MTA map difficult to use.

Figure 5-3. The 2004 version of the MTA New York City subway map, based on a design by Michael Hertz. Besides its visual complexity, incomplete information missing on the map itself forces the the user to rely on the complex charts in the lower right section — right where sitting people block its view in the subway cars — and in the stations where this information, displayed on large posters, is also difficult to read since it is often less than 18 inches off the ground. New York City Subway Map © Metropolitan transportation authority. Used with permission.

It was one thing to reduce the size, but another thing to realize that the way the data was presented was not the best way to present it. So I asked myself: how would I present all this data?

To answer this question, I had to ask a few more:

  • What maps came before this map?
  • Were there any previous conceptions that were discarded but perhaps still relevant?
  • What was it about New York City and its subway that historically made it so difficult to map clearly and efficiently?

Looking back to look forward

I did a research dive, and I started buying old transit maps on eBay. I studied subway maps, New York City street maps, and transit maps from all over the world that I had collected on my travels. I filtered through all the design approaches and eclectically took as much as I could from ideas that had already been implemented (some brilliantly).

Of course, in addition to the map designed by George Salomon that had been my father’s subway map, I studied carefully the map designed by Massimo Vignelli (see Figure 5-4), which the MTA used from 1972 until 1979, when it was replaced by the Tauranac-Hertz MTA map (which, 30 years later, still prevails). Vignelli’s map appealed to me immediately because, although big, it took obvious inspiration from Beck’s Tube map, with its 90- and 45-degree angles, explicit station connections, and the use of color to denote individual lines. There were also some smart aspects of the current MTA map that I wanted to keep, despite finding it on the whole unwieldy because there is so much information crammed onto it. In addition, I borrowed liberally from other past efforts that had been discarded or forgotten.

Figure 5-4. The 1972 MTA New York City subway map designed by Massimo Vignelli. Confusingly distorted geography for style’s sake — yet a stunning design icon. 1972 New York City Subway Map © MTA New York City Transit. Used with permission.

New York’s unique complexity

As I conducted my research, I started to realize that New York City had its own unique set of challenges that made its subway system impossible to accurately and clearly map using just a diagrammatic method, as other cities like London, Paris, and Tokyo had done. It was also clear that a pure topographic mapping approach wouldn’t work, either; New York’s unique geography and its gridiron street system both have an impact on mapping its subway system.

There are four significant and conflicting aspects of the New York City subway system that make it impossible to successfully map with either a strict diagrammatic or topographic format:

  • The narrow geography of the principal thoroughfare, Manhattan Island, which has 17 separate subway lines running up and down Midtown alone in a width of six city blocks.
  • The “cut and cover” method used to construct subway tunnels and elevated lines that follow the city’s gridiron street patterns. Because New York City’s subway generally follows its gridded street routes, there is a strong psychological link between the subway and the aboveground topography that is not found in a medieval city like London.
  • The unique system of many of the subway lines running local, then express, then local again along their routes.
  • Its formative history, with the current system evolving from three separate and competing subway systems (the IRT, BMT, and IND) that were poorly coordinated to work as a whole system. (The chaotic tangle of these three competing routes, as they meander and fight their way through the dense street plans of lower Manhattan, downtown Brooklyn, and Long Island City, is the most difficult part of the system to map clearly and accurately.)

The KickMap, shown in Figure 5-5, is based on a combination of ideas I selectively borrowed from many earlier maps (some dating back to the 19th century) and my own innovations. I believe that this unique combination makes my map easier to use than most of the preceding efforts. In the following sections, I’ll discuss my inspirations and innovations in more detail.

Figure 5-5. The KickMap as it was released in 2007.

Geography is about relationships

Most of the boroughs — Queens, Brooklyn, Manhattan, and to some extent, the Bronx — already have a grid on top of the subway system because of the way the streets were planned. This makes the aboveground geography not only an intuitive starting point, but also an integral part of the user’s experience. Knowing your location — take 42nd Street and 7th Avenue as an example — places you in the grid, which makes it easy to judge distances and locations. This is why the numerous geographical errors that appear in New York City subway maps (like the Vignelli map infamously placing the 50th Street and Broadway stop west of 8th Avenue instead of east) are so glaring and easy to spot.

One of the issues I have with some previous versions of the New York subway map is that I have a hard time believing that the designers ever actually rode the subway as an integral part of their lives in the city. There’s a disconnect between many of the decisions they made and the reality of the subway. As part of my design process, I rode the lines and exited the stations at every major intersection with which I was unfamiliar. There is a strong relationship in New York between the aboveground and the belowground, and since subway riders don’t cease to exist when they leave the subway, it’s important for the map to express this relationship as clearly as possible. Otherwise, the result is an uncomfortable feeling of disorientation.

Include the essentials

Consider the L line in Brooklyn. As a passenger on the train, you’re jostled around as you travel and you don’t really notice that the line is curving or turning corners along major streets and intersections. But when you get out at the Graham Avenue station, for instance, it’s obvious that Metropolitan Avenue and Bushwick Avenue are two major thoroughfares that intersect each other at a right angle. Why wouldn’t that show up on the map? If you didn’t know how the streets intersected and you just saw a sign for one or the other as you came out of the subway, it would be very difficult to figure out what was going on.

On the Vignelli map, this portion of the L is depicted as a straight line (see Figure 5-6[a]). The Hertz map (Figure 5-6[c]) shows both Metropolitan and Bushwick Avenues, but the line resembles nothing so much as a wet noodle as it half-heartedly depicts the route. I chose to carefully draw a stylized but accurate line describing the path as it runs along each major avenue there, believing this to be the best approach because it is the most helpful to riders (Figure 5-6[b]).

Figure 5-6. A portion of the L line in Brooklyn as depicted by (a) the Vignelli map, (b) the KickMap, and (c) the Tauranac-Hertz map.

Conversely, I sometimes made stylistic simplifications to the geography in order to help riders. For example, Queens Boulevard, a major thoroughfare in Queens, was originally five different farm roads, and as a result it jigs and jogs a bit as it makes its way from the Queensboro Bridge east across the borough. Recent maps didn’t capture its relationship to the subway because they either ignored it entirely (as in the Vignelli map, shown in Figure 5-7[a]) or obscured it (as in the current MTA map, shown in Figure 5-7[c]).

On my map, I styled Queens Boulevard as a straight line; see Figure 5-7(b). I chose to do this so that users could easily see its path and identify the “trade-off” subway lines that travel along it — where one subway line runs along the road and then veers off and another line takes its place. In this case, the 7 line runs along Queens Boulevard until it veers off along Roosevelt Avenue, and the R/V/G/E/F lines come down from Broadway and pick up its path east. My stylized approach uses logic to better convey the subway’s relationship to the streets of Queens, which is not clearly apparent on either the Vignelli map or the current MTA map.

Figure 5-7. The trade-off along Queens Boulevard as depicted by (a) the Vignelli map, (b) the KickMap, and (c) the current MTA map.

Another “trade-off” I felt it was important to show clearly is at 42nd Street in Midtown Manhattan, where the 4/5/6 line jogs over from Park Avenue to Lexington Avenue (see Figure 5-8). A would-be rider walking along in Midtown or Murray Hill needs to know which street to go to for a subway entrance. The Vignelli map obscures the shift by treating it as a straight line, relying on text to convey the road switch, and once again the current MTA map is at best vague and noodley. In my map, it’s clear which way the user should go.

Figure 5-8. A portion of the 4/5/6 line in Manhattan as depicted by (a) the Vignelli map, (b) the KickMap, and (c) the current MTA map.

Leave out the clutter

While I felt that it was important to show certain shapes aboveground, I also felt that it was important to leave out certain pieces of belowground information. There are several places where the subway tunnels cross and overlap each other beneath the surface. This may be important information for city workers or utility companies trying to make repairs, but for the average commuter, showing these interactions just creates visual noise. I tried to reduce that noise by cleanly separating the lines on the map so they don’t overlap. Consider the different depictions of the 4 line and the 5 line in the Bronx (Figure 5-9); sure, the MTA’s paths may be accurate, but they’re also confusing, and riders don’t really need to see those particular details to understand where they’re going.

Figure 5-9. The 4 line and the 5 line as depicted by (a) the KickMap and (b) the current MTA map.

Coloring Inside the lines

The belowground geography is important, but it’s more vital for the users to understand which belowground lines will take them where they want to go.

In 1967, the MTA moved past the tricolor theme used on the Salomon and earlier maps and began to use individual colors to illustrate individual lines. However, this shift didn’t help simplify the system. It essentially had 26 lines assigned 26 random colors, which didn’t really tell the user anything beyond illustrating the continuity of a given route. Vignelli’s map (Figure 5-10[a]) continued with this color system.

The Tauranac-Hertz (current MTA) map attempted to simplify things by collapsing multiple subway lines onto one graphic line, but this actually made understanding the subway system more complicated, as now you had to read the text next to each and every station to learn whether a specific line stopped there or not; see Figure 5-10(c). What it did get right was that it color-coded sets of subway lines that use the same track — for example, the A/C/E lines are all blue, and the 4/5/6 lines are all green. If you look at the “trunk” lines that run north and south through Manhattan, the colors move from blue to red to orange to yellow to green, creating a spectrum effect. These colors are memorable and help riders discern which lines will take them in the general direction they want to go.

In my map, I preserved the best elements of both approaches; see Figure 5-10(b). I reused the spectral colors on the trunk lines, highlighting an elegance and reality inherent in the system that Tauranac-Hertz understood, but kept it clear by representing each route with its own graphic line. Technically, I did what Vignelli did in that I used 26 distinct colors, but I grouped them in six or seven families of color and used different shades for each line in a given family: the A/C/E lines use shades of blue, the 4/5/6 lines use shades of green, and so on.

Figure 5-10. The Manhattan “trunk” lines as depicted by (a) the current MTA map, (b) the KickMap, and (c) the Vignelli map.

I also made use of line IDs and colors for the station dots (2). The crucial idea here was that the map should be quickly scannable, rather than just readable. At each station where a line stops, I placed the name of that line inside a dot: this way, users can easily see exactly which trains stop at which stations without having to read a list of lines next to each station name. Use of different colored dots enables users to tell at a glance whether the train always stops there or has special conditions, such as weekday/weekend or peak hour/off-peak hour restrictions.

Finally, there are about 80 stations in the city where, if you’ve missed your stop, you can’t just get out and conveniently switch direction. I highlighted these locations by placing a small red square next to the station name, indicating to riders who need to turn around which stations to avoid if they don’t want to have to leave the station, cross the street, and re-enter the station on the opposite side. The current MTA map shows all the heliports in the city but doesn’t provide users with this simple but important piece of subway information — a perfect example of its confused priorities.

I believe that taken together, these decisions highlight the innovations that make the KickMap more useable than those that came before it.

Sweat the small stuff

Those decisions were easy for me, but other choices were more difficult. Which geographic features did I really need to keep? What angles should I use? How much bus and ferry information should I include?

So, after creating my first comprehensive map that met my initial challenges (Figure 5-5), I decided to refine it and incorporate all of my learning. I was excited.

Try it on

In the car industry, it is common to build what is called a test mule, which is a prototype or preproduction car into which every possible experimental feature is crammed; that prototype then undergoes a series of drivability tests to determine what should be removed (because it’s not essential or doesn’t work quite right). I did the same thing with my map: I created a version (shown in Figure 5-11) into which I put every feature that I might possibly want. Illustrator’s layers feature really came in handy here; I put a lot in this map that I ultimately turned off or toned down.

Figure 5-11. My version of a test mule for the map: I put lots of information in and then edited it down.

The mule map allowed me to evaluate a variety of trade-offs, such as:

The street grid — I wanted to present the structure of the streets without interfering with the subway info wherever I could. You’ll notice that the mule map includes a lot more streets and street names than the final design.

Beaches — I thought green spaces were important, and that New Yorkers should be able to find their way to beaches by subway rather than by car. My mule map included municipal swimming pools as well, but ultimately I decided to remove them.

Coastline features — It was important that real people — like, say, my mom — could easily use this map, and she couldn’t care less about certain geographic details (like Steinway Creek or Wallabout Bay) that I included in the mule map. That was a reason to simplify and stylize. But I also wanted to make something any map geek or lover of New York City (like me!) could appreciate. So, there were instances where I let my passion take over. I decided to pay homage to certain subway feats, so I included features like the Gowanus Canal, which the Smith/9th Street station crosses and has to clear (at 91 feet, it’s the highest elevated station in the system).

Angular design — In the final design I standardized a lot of the angles, but I broke that standardization if I had to for clarity’s sake. I wasn’t a slave to the angles. Stylization is fine, but my goal was to take the stylization and make it work so that riders can always understand what’s going on aboveground. I also decided to consistently place station names on the horizontal for easier reading, like on the London Tube map, instead of cramming them in at arbitrary angles.

Bridges and tunnels — One of my goals for this project was to come up with a tool that would encourage people to take the subway instead of a car. For this reason, I decided to leave out all the car bridges and tunnels (except for the iconic Brooklyn Bridge). I wanted to keep the experience of navigating the subway as clean and easy as possible, without the temptation of using a car, to encourage users to keep riding.

Many of these choices were influenced by the following principle.

Users are only human

There are certain New York icons that help orient the rider and are reassuring. To the extent that they represent something familiar, maps can be quite emotional. So, I saw preserving such icons as a way to build friendliness into this tool. I did not design a geographically precise topographical map; I designed a map that is emotionally and geographically accurate in a relational sense — Manhattan looks like Manhattan, Central Park is green, the Hudson River is blue, and the subway stations are positionally accurate in relation to one another and the streets (Delancey Street is shown east of the Bowery, etc.).

For the same humanistic reason, I included certain celebrated landmarks — the Statue of Liberty, Ellis Island, and the Brooklyn Bridge. And I didn’t just include them with nametags; I actually included their familiar shapes, as was done on subway maps back in the 1930s (3).

A city of neighborhoods

When I travel on the subway to see my mom, I’m not going to see her at the 95th Street subway station; I’m going to see her at her home, which is in Bay Ridge, Brooklyn. This is an important aspect of New York: it is a city made up of neighborhoods, and native New Yorkers think of the city in those terms. That’s our frame of reference: we travel from, say, Washington Heights to Bay Ridge.

The current MTA subway map includes some neighborhood names, but they are just dark blue words that compete with the station names and do little to describe the areas. There’s no hierarchy of information. By color-coding the neighborhoods — which has been done on maps of the city since at least the 1840s — in an unobtrusive way (using pastel tones) and writing their labels in white text so they wouldn’t visually interfere with the black text of the station names, I was able to provide layers of information without compromising the clarity and functionality of the subway map.

Again, these elements were literally created in separate digital layers in Illustrator. This allowed me to turn the neighborhoods on and off to determine what really needed to be there and to make several variations of the subway map with and without them.

One size does not fit all

I believe that separating functions is an important key to any useful visualization or tool.

Another benefit of the layered approach was that it allowed me to custom-tailor the map to the user interface later. The KickMap is available as iPhone and iPad applications, and in that context, the map’s detail automatically changes as the user zooms in or out. Besides the apps, commuters still read subway maps in many different contexts: there is the foldout printed version, the huge ones they hang in the stations, the ones they post in the train cars (right behind the seats so that you have to peer past someone’s ear to read them), and the one that is posted online. Currently, you get basically the same map in each place, but that shouldn’t be the case: in each context, a slightly different version, optimized just for that specific environment, should be available.

Each version should have its own design, tailored to the context in which it appears. The big maps that hang in the stations, for instance, should show you the neighborhoods, but the one in the subway car that riders reference to make quick decisions, like whether to get off at the next station, need not. And why does the map in the subway car have to give you all that bus information?

Contexts aren’t just physical, either. After 11:00 pm in New York, 26 routes reduce to 19. So, in addition to the main day/evening KickMap, I made the night map shown in Figure 5-12. Instead of relying on a text-heavy, hard-to-read chart at the bottom of a one-size-fits-all map to determine when a certain route is available, a night map should be available to riders (not only on their iPhones, but also in the subway cars).

Figure 5-12. The night version of the KickMap shows only the lines that run between 11:00 p.m. and 6:30 a.m.

When it came to making a night map, I simplified the day/evening version and took out most of the street and neighborhood information, as it seemed redundant.

Also, I do love the simple and elegant aesthetic of Beck’s Underground map, and keeping the night map’s form simple pays homage to it!


Ultimately, I do think the KickMap accomplished most of my goals: to make the subway lines and their connections as clear as possible for easier navigation, and to provide users with a clear representation of where they are once they exit a station so that the subway feels familiar and welcoming to all.

My main goal, however, was to get my map out there into the hands of subway riders. After the MTA rejected my design, I found an alternative way to distribute it, via Apple’s iTunes — two apps, one free and one paid, for the iPhone, iPod Touch, and iPad.

All of the choices I made were aimed at trying to make the user experience as seamless and pleasant as possible. Clearly I’m striking a chord, as over 250,000 people (and counting) have now downloaded copies of the KickMap from iTunes. That’s really great but I still want the KickMap — or something superior — to replace the current one in the subway system. I want people to be comfortable and even happy when using our unbeatable 24-hour subway system. It is a complex system, but if people know how easy it can be — if the map becomes a friend (4) instead of an obstacle — ridership will increase. Ultimately, that benefits not only the system itself, but also all of us who live, work, visit, and breathe here.



1. I now know that map was an early version of the Salomon map. Years later, when I was doing research for the creation of the KickMap, I got to appreciate the beauty of the design of this map.

2. This was a big aha moment in my process.

3. I wanted to put the Empire State Building in there, but it would have cluttered up Midtown, and my goal all along was that it really had to be a simple and functional subway map!

4. I think many people are passionate about the subway map as a great symbol of New York. The map shows the subway as kind of a dynamic capillary system nourishing the city. This is true not only conceptually but also historically: the subway was built to “nourish” new residential areas with cheap transportation to and from the central business districts so the City could continue to grow and thrive.

Post topics: Data