Information Pollution on the Underground Map


I'm going to start by saying something obvious, but bear with me, because if it really was so obvious, I wouldn't have to write this piece.


Humans have limited cognitive capacity. The more information they are given, the more likely they will make a mistake making use of it. Even if the information is of peripheral importance to the task in hand, it can still get in the way and distract from it. Any given task has a cognitive load associated with it. The higher the cognitive load, the more cognitive capacity required to cope adequately.


Humans differ in their cognitive capacity. Some people are better able to cope with large quantities of information than others. Cognitive capacity is related to intelligence, a dirty word in many quarters, but nonetheless an important one. Tasks with a higher cognitive load are more of a challenge for people with lower cognitive capacity. Politically correct people who dismiss the importance of intelligence may unwittingly discriminate against people with low cognitive capacity. Not so politically correct after all.


For providers of information, they have a responsibility to reduce the cognitive load wherever possible. They need to give people exactly the information that most of them will need to complete a task, and no more. Providers should structure and break down tasks so that people don't have to receive all relevant information all at once, only exactly what is needed for each stage of the decision making process. Providers should ensure that information is clear and unambiguous. A flight of steps is a barrier to a person who is mobility impaired, poor provision of information is a barrier for a person who is cognitively impaired.


The world is a complex place, so how do we cope with all the information that bombards us from every direction? Much of it is of no use for the current task, so we try to screen it out. To do this successfully requires cognitive capacity, but not as much capacity as trying to make use of all the information. Sometimes we screen out too much information, missing something important. This is why repetitive trivial station and on-train announcements are at best irritating, and at worst dangerous; having got us into the habit of ignoring them, genuinely important messages may be screened out too.


Our chief weapon in defeating information overload is expertise and experience. This helps us decide in advance what information is important and what is not, so that screening out useless information is easier. Frequent tasks can be performed automatically, permitting scarce cognitive capacity to be devoted to fathoming out what is out of the ordinary, where experience can offer less help. People with the greatest cognitive capacity can make inferences from their past experience, for example using analogies to apply old knowledge to new situations. High cognitive capacity better enables us to acquire expertise and capitalise on our experience.


This all seems very sensible, but what has it got to do with maps? Our problem is that there is a creeping malaise in the bureaucracy business, on the one hand a belief that if one piece of information is helpful, ten pieces of information will be ten times as helpful, and on the other hand a fear that if people are not told everything that might possibly be relevant no matter how peripheral to the task, then something terrible might happen. This has led to a sort of feature creep, where, amongst other things, maps become more and more overloaded with details, and harder and harder to use, especially for people who are novices or have low cognitive capacity. If these details are not really necessary, and just get in the way, then we have information pollution.


What counts as information pollution exactly?


The status of some information on the map isn't clear, because it is useful for some people but not others. Zones are a good example. Their application to the current map is extremely clumsy, making lines harder to follow and stations harder to identify. When printed with too dark a tint, they result in uncomfortable stripy effects, when printed too light, it can be difficult to identify which zone a station is in. Even if the information itself is not pollution, its current means of application has undoubtedly polluted the map.



Before and after (I)



Before and after (II)



Before and after (III)



Before and after (IV)

Applying zonal information to the map has not been cost-free in terms of usability. For a person with an all-zones travelcard, or a fully-charged Oyster pay-as-you-go card, zones on the map are information pollution. For a person with a zone-restricted travelcard, this information is more important, although it is also given on the station index on the reverse of the map (itself poorly designed and with tiny lettering, sponsorship appears to me more important than usability), in stations, and on car line stripmaps. For what percentage of travellers is zone information on the map essential, for what proportion is it information pollution?


Zones add to the quantity of information that must be navigated, and many travellers will need to screen this out, but we can at least envisage other travellers wanting to make use of it. Other information adds to the quantity without obviously being helpful in any way. Remember that an Underground map is supposed to be a journey planner, how do I best get from A to B? The January 2008 journey planner warns me that Tower Gateway will be closed in the summer (there is no mention of the Bank interchange closure in March), that Wood Lane station will be open late in 2008, that in 2010 New Cross Gate to Crystal Palace and West Croydon will join the Overground Network, but that Whitechapel to Highbury will join it at an unspecified date. None of this information helped me to plan a journey in January 2008. Some of it is well-intentioned, the rest is just corporate vanity, apart from possibly helping me to decide where to buy a house in 2010.


Information on the Underground map is only essential if it has a direct bearing on the journey planning process in some way. Does it affect the start station? The end station? The route between the two? If not, then on-station or on-train instructions might be better. Remember, break down the task into manageable chunks, feed people information as they need it, not all at once.



Tower Gateway is closed in June, but I want to make my journey in April



How many people travelling to/from Mill Hill are going to change their journey plans because through trains might not be running?



This is how the East London Line closure was handled in 1996, compared with 2008. It is helpful to know in advance that Wapping to Rotherhithe is impossible, but do I really need to be faced with all those different bus routes before I get to an ELL station?

Not just the quantity of information, but also the quality


High quantity of information makes life harder, but poor quality of information makes life harder still, it disproportionately increases cognitive load. Here are some examples of how:

A nice example of poor quality information on the current map is the warning daggers used to denote various caveats with stations and lines. Not only are there very many of them, but they also vary in importance from the utterly trivial to the extremely important. Worse still, many have been thoughtlessly, inconsistently, or incorrectly applied. Here are some examples from the February 2008 card folder:

Daggers have long been used to flag additional service details, but a new development in January adds yet more inconsistencies to the map. The next four images flag interchanges in three different ways. Does this mean that there are three different types of interchange?



Three different ways of showing something identical: an interchange between lines which involves leaving one station, and hunting around in local streets to find the other. What is the user to make of this?

One more example of inconsistency/incongurency on the map. This concerns the appearance of lines themselves. These examples are all from the June 2006 card folder, one of the worst Underground maps ever in terms of information design.



In the past, 'pecking' has traditionally denoted a section of line with service restrictions. It still does sometimes (for example, compare Kennington to Oval with Waterloo to Bank on any recent map), but is taken to an extreme in this example. On one line, the pecking means 'no service except during rush hours', on the other line, the pecking means 'no service for three months'.



What about casing then? In the past, it has usually indicated a service of interest to Underground users, but not actually run by the Underground. It sort of still means this, with the DLR and Overground cased, but in June 2006, it acquired a second meaning: a closed section of Underground line with a replacement bus service.



OK, so when a line is closed, pecking means no replacement bus service, casing means that a replacement bus service is offered. Got that?



Sorry, false alarm. Pecked closed lines can have a replacement bus service too!


This is quite a long list of problems, and I hope you have kept with me so far. There is a very high quantity of information on the current Underground map and, worryingly, much of this is poor quality. The odd glitch here and there might not matter, but when these sorts of problems stack up then it is inevitable that cognitive load will increase and usability will suffer.


You can't help the user without understanding the user


How does this help? If only this question was asked more often, we would be spared so much half-useful information that is worse than useless. It clutters things up without actually telling us what we really need to know. Take British Rail interchanges on the Underground map. Knowing that an interchange is available with an invisible service is not helpful until we know which exact option is the correct one. If we know which is the correct option, then we don't need the extra help from the map.



Going to Brighton without a timetable or knowing which mainline station to travel from? Good luck! I make it 13 possible stations to choose from in this section of the map alone, only four of these will get you directly to Brighton. If you knew which mainline station you needed, why did you need this information to be duplicated on the map?

Now that aircraft have been added too, we have yet more half-useful information. "Interchange with National Rail services to airport" Yes, but which one? Any tourist who thinks that these all refer to Heathrow is going to get into trouble. We have riverboats too to add to complexity, although most of the services are just guided tours. At least there is currently only one station that links to Tramlink, if a user must get to the Tramlink network, there is apparently only one way to do it. What next? Buses? Shops? Every Underground station interchanges with something else, why not just replace all station tickmarks with interchange circles?



According to the March 2008 edition of the Underground map, all these stations can get you to an airport, but which airport? Again, if you know the answer, why do you need the map to remind you?

Those wheelchair blobs


The addition of wheelchair blobs to the map in June 2006 to denote step-free access from street to platform caused a great deal of controversy. The problem with them is that they are simplistic and misleading. No person who genuinely required step-free access to the Underground could make use of this information as presented in this way on the map. This means that wheelchair blobs count as information pollution for everyone using the map (wheelchair users should not be relying on the standard underground map under any circumstances, there is a comprehensive access guide for explaining these complexities). Here's why:


Step-free access from street to platform for three stations in Northwest London? Unfortunately, these platforms are 'compromise height', served by differently-sized trains, so it is not necessarily step-free access from street to train.

Step-free access from street to platform for certain lines in Central London, but can you tell which entrance from the map?

Here are a selection of level (i.e. no lifts necessary) interchanges currently available to Underground users, either between lines or services, or with national rail services. Only interchanges in both directions are shown (e.g. northbound with northbound, southbound with southbound), a number of single-direction interchanges are also available. The standard Underground map does not even hint at their availability.

To get from Woodford to Westminster, don't change to the Jubilee Line at Stratford (lift), even though this route is implied by the map. Change to the District Line at Mile End (cross platform), except that the interchange at Mile End is not shown anyway.

So, the wheelchair blobs are there for box ticking, not much else (and don't forget that many people with mobility problems can cope with a couple of stairs or an escalator, but may need to avoid a long walk). Given their low utility, their cost need not be too great to outweigh the benefits of including them.


Quick! Which of these stations are interchanges, which are not?

Having one important symbol obliterated by another is particularly poor information design. Either interchange circles are important, then they should always be shown, or they are not important, and so they should never be shown.



Is the user recommended to change between stations at Hammersmith, or not?

So what should be done?


There seems to be an assumption made by many people that all a designer has to do is create a map using the same sorts of rules that Henry Beck did 75 years ago, and a design masterpiece will magically appear. This is not remotely true, and even if the geometry of the map is impeccable, serious usability damage will result if excessive poor quality information is added.



Ideally, a map of a rail network should show just: (I) stations; (II) lines; and (III) opportunities for changing between lines. Here is the March 2008 Underground map showing everything that has been added over and above this (readers may wish to compare the quantity with other networks such as Paris or Madrid). Note how the effect of this by itself is to draw attention away from the centre and towards the southeast. All this information is potentially distracting, so it had better be important for most users, and of high quality. If not, then the smog of information pollution will interfere with journey planning.

If the Underground map is an important navigation tool, then it is important that it is an effective navigation tool for all users, no matter what their level of experience or their cognitive capacity. All of the difficulties raised above are fixable. Controlling the level of information pollution on the Underground map (both quantity and quality) should not be difficult nor time consuming, nor expensive to do. Of course, many of the solutions require an integrated approach to be taken to journey planning, signage and system usage, so that the various tasks that must be performed in order to plan and implement a journey are understood, and the information provision structured accordingly. From the point of the view of the map though, the key is to help the user by understanding the user.


Last updated 23/06/08. Click to get back to my home page