In the continuing fight for the greater availability of public information, it may seem churlish to observe that sometimes what’s wanted is not more information, but less.
The picture above shows a typical display on a Countdown sign at a London bus stop. This particular stop has buses from two routes. At a quick glance, you might be forgiven for thinking that it’s going to be a long wait for a 2, while three 88s roll past ahead of it. Actually, the sign shows no such thing: the bottom three lines cycle through all the buses known to the system, so the second, third and fourth aren’t showing. Nor indeed are the eighth and ninth, which at that moment were 20 and 23 minutes away respectively. But if you want to know when the next 2 is due, you will have to wait for the sign to cycle round.
I would be prepared to lay considerable odds that there is not a single person standing at that bus stop who has the slightest interest in the fact that a 2 will be along in 16 minutes, still less (if that were possible) that there will be another one in 20 minutes. What they care about is that the next one is three minutes away, and right now the sign isn’t telling them that. In extreme cases, there are so many buses stacked up in the queue that they come and go faster than the display can keep up with them – the screenshot on the right (from James Darling’s minimalist site) is from the bus stop with the most extreme case of that problem I have come across.
There are no imaginable circumstances when it would be useful to know that that the sixth bus to Marylebone is coming in 26 minutes, when you are already standing at the bus stop. But there is potentially much greater value in that information when you are not at the bus stop at all. So for Countdown at bus stops, perhaps the solution is simply to show less. The first and second bus due for each route served by the stop might do it (though capping by either time or an arbitrary number of buses shown wouldn’t).
But the moment you are not at a bus stop, the ideal solution changes radically, for two reasons. The first is that the information is now informing a different decision: whether now is a good time to be going to the bus stop in the first place, with opportunities for trade offs which exist only because the information has been liberated from the bus stop. The second is that the small screen in my pocket can comfortably show me the next twenty buses; the much bigger screen attached to the bus stop struggles as soon as there are more than four.
The point of this is one which at one level should be trivially self-evident: what counts as good information depends on the question you want to answer or the problem you want to solve, and more information may not be better information. But since there are many different questions and problems, it would be folly to think that there is a single best way of selecting and presenting the information. Part of the power of open data is that it creates the possibility for highly specialised solutions – as Adrian Short has both elegantly argued and practically demonstrated. James Darling’s account of why he built his version and what he and others have done with it is also well worth reading.
One of the comments to that post links to a more radical approach still: stop making it about the buses (tubes, trains…) at all, and turn it round to be about the passenger, in a form of extreme hyperlocalism:
As the about page says, if you live exactly 6 minutes from Sunset Tunnel East Portal, 8 minutes from Duboce and Church, and 10 minutes from Church Station you may find it useful too.
But enough. I have a bus to catch.
I had hesitated to write this post at all – there is only so much bus stop nerdery any self-respecting blog should contain, and I am already well over quota for this year. But I took heart from Paul Annett’s meticulous deconstruction of a bus stop indicator at Heathrow, so here it is.
Stefan, you make some really good points. When I lived in London I thought listing every single bus at the stop was completely pointless, because clearly I was going to get on the next arriving bus. The only time I appreciated all of that information was if I was picking someone up at the bus stop – then I could see when all the possible times the buses were arriving, and could determine whether I should wait for them in the rain or go grab a coffee 🙂
As E. Tufte is always saying, it isn’t the quantity of information you are providing that is the problem, it is the way you offer it. What we have here is TOO LITTLE of the right information. A different way of displaying, a better selection of data, or a more graphical display is likely all that is needed.
Better decisions are almost always made by considering more information. However, as the research literature on expertise clearly tells us, one of the things that experts in any field are able to do is filter out what they immediately recognize as irrelevant information. That’s part of how they reach expert decisions so efficiently despite having more knowledge to draw on. Their expertise allows them to actively reduce the amount of information to be mentally juggled at that moment.
We see this illustrated most effectively in professional sports where athletes astound us with their capacity to make snap decisions, whether it is a hockey player’s ability to be at the right spot at the right time to slap a rebound into the net, a running back’s ability to “see” a hole open up and run a breaking pattern that avoids potential tackles, or a NASCAR driver’s ability to identify a potential opening at ridiculous speeds. They can do this because they incorporate more into their decision, but saddle themselves with less to think about at any given moment. They have the right stuff, because they are thinking about the right stuff.
And that’s the objective: how to provide people with the full spectrum of information they might need, but allow they to prioritize and ignore large swathes of it.
One of the things I am often complimented on during presentations and inpapers are the figures and diagrams, that allow people to bring a lot into how they think about something, let them get perspective, so as to develop higher-order insights, yet not drown them in details.
A friend and neighbour of mine works in the area of “human factors”, primarily with respect to displays, dashboards, and such. His focus is on allowing people to derive as much well-informed understanding as possible at a glance. Currently he’s working on operating-room info systems, such that anesthesiologists can look up at a “dashboard” of multiple metabolic functions and know where the patient stands at this moment, with higher-order perspective and without confusion. Happily, there’s a lot of people engaged in research like that. Unfortunately, they aren’t always consulted, and what they know is not always fruitfully implemented
But I am not sure how far any of that applies to my humble bus stop example. My perception there is that information of no value is crowding out information of high value. The simplest – and most elegant – solution to that is simply to remove the stuff which serves only to get in the way.
So to that extent (and I entirely accept, only to that extent) less really is more, and the usefulness of the information would be enhanced by recognizing that.
The examples you provide are prime examples of what happens when the information displayed is determined by the folks who do the data-management – the programmers – and not determined by the end-user. They will do their very best, I’m sure, but their inclination is always to go towards the easiest implementation from a programming and data-access point of view, and not from a user-comprehension point of view. One of the many reasons why my neighbour teaches a course on such matters to engineers.
I used to teach what was sort of an everything-you-need-to-know-about-applied-psychology course to computer science folks. Tough slogging, I have to say. One of the modules I taught was on documentation writing for user-comprehension purposes. As an exercise, I had the class collaboratively write documentation for something they knew extremely well, but directed at someone who knew very little about it. One year we did documentation for Monopoly, and the following year we did Scrabble. Their inclination was to throw anything and everything at the user, and rush to the fancy complicated stuff. I would have to keep asking them “Does the novice want or need to know this now?” and “What question do you think is running through their mind at this point?”. For a variety of reasons I won’t digress into, there is a kind of universal tradeoff between the social/verbal and the spatial/quantitative, when it comes to human abilities, and folks in that line of work are not as likely to be competent at that sort of perspective taking. Needless to say, after 90 minutes, their hair was matted to their foreheads, and they were exhausted. What they thought was going to be a cakewalk had pushed them well beyond the perimeter of their expertise.
My point is that information furnishers, and information users can often operate in two entirely different spheres.
At the moment, I’m preparing a summary report of some large-scale survey results, directed at folks who would a) simply like some confirmation anmd feedback about the survey they participated in, and b) sincerely want to know what the “state of staffing is this year”. I’m neck deep in numbers, and have a near infinity of fascinating breakouts, crosstabs and such that I could present. My task is to convey as much as possible in a ways that:
1) lets them know I have their priorities in mind (the face validity mission)
2) lets them encounter the quantitative information at a reasonable pace (small bite-size mouthfuls)
3) lets them identify the important results easily
4) lets them walk away with a narrative that is relatively easy to commit to memory, without being overly simplistic, so that one reader can easily translate into 3 or 4 “knowers” (i.e., people who didn’t read the report but get a faithful transmission 2nd hand from a reader)
While your initial examples focus on something as seemingly trivial as a bus arrival schedule, I think the same principles apply to that as to report writing.
The short versus the long answer. There’s a place for both. I agree, Mark, that all the information isn’t required at once. Now, we say, give us a link; we used to say give us a handout. Also, now we have to download the instruction manual that may instruct us how to start-up our new electronic toy–oops. It’s that just enough feature that has everyone baffled. Write anything and everything for the user. Some users need a lot of data before they are satisfied, some just want to know you know it, and others don’t care so if you must, give them enough of an outline that they’ll be satisfied.