Skip main navigation

Crowdsourcing a sense of the city

Daniele Quercia explains what data from public transport smart cards and a game built on Google Maps can tell us about life in the city.
Most of my research has been around the stunning psychological perception of cities. So when you are in the built environment or when you are in a certain neighbourhood, how do you feel without the psychological perception of it? And when we started, we started in a very quantitative way, and as some people in the room, we started to analyse the Oyster card data. So you know what it is, right? You do. Most of you are in the UK so I assume that– how many people use the Oyster card? Use, use, use? Oh, a lot of people. Fantastic. So it’s basically a passive RFID, if you scratch it. I don’t know if you did it.
But if you scratch it, then you can find that. And that means that, basically, when you tap in, it’s registering that you’re entering a station. When you tap out, you’re exiting the station. Now if you collectively take the data you can actually study the flow of people. And that’s what we did for a month. So for a month, we studied the flow, data tapping in and tapping out. And we studied a graph of the London Underground where you had nodes of the subway stations and the flow of people going from one subway station and the others. And we studied some specific metrics that are related to centrality or between-ness in a network.
But if we were to summarise really quickly, the main finding is this one. This is science at best. So not really. So what we found is that people who live, who live in deprived areas, they go to other deprived areas, and they go to well-off areas. People who live in well-off areas, what do they do?
They go to other well-off areas. So that’s the obvious thing we found. And one sociologist– how many sociologists we have in the room?
Oh, wow. Cool. Welcome, guys.
So was a show that BBC radio said, oh, that’s obvious, right? We’ve been knowing that for 30 years. But they conceded that we’d be knowing that now from different data sources, real time and different temporal and geographic resolutions. But what we know is that the concept of visibility. Right? So this is well-known. Most of you might know that. And the idea is that well-off areas, they tend to be much more exposed than deprived areas. And then they tend to stick in people’s mind. First guy, Kevin Lynch, some of you, especially the urban planners, architects, in the room have read his book. 1970, he tried to study different city forms and tried to relate that to the well-being of city residents.
How does he define visibility in a very, very easy way? So if you can organise in a coherent pattern different parts of the city, then that part of the city tends to be visible. How that relates to well-being though, it’s a bit more complicated. And so, just let me give you an example. The idea is that if you were to leave– I was speaking with someone yesterday night and I gave this example. Just imagine you have a house, and from your house, you get out and you are in a maze, right? So you don’t find your way. So every day you have to walk through how – expensive in cognitive terms, is that, walking through a maze?
Now think about another situation where you were to live in a very distinctive part of the city, easy navigable. So you take always the same route. Sometimes you can change and that is easy to do, right? So in terms of mental effort, it’s easy. So there is a relationship between city forms and well-being.
Have you been following me? OK, fantastic. So city relationship, Kevin Lynch, 1970s, started to hypothesise that, but the first quantitative study, between this relationship between well-being and city forms is that of Milgram. How many people know Milgram? Tonnes of you. Most of you. Most of the computer scientists know him because of the six degree experiment. But he did other experiments. And one thing that he did is how can I quantify the visibility, right? I gave you the definition before, but it’s very difficult to quantify that. So what he did, he did a very simple experiment. And he said, OK, a good proxy for visibility is recognisability.
So what he did is that he had his students in front of him in New York back then. And he showed them for 90 minutes different pictures of New York. And the student had to recognise where the picture was taken. So after collecting all answers from his students, and measuring the fraction of correct answers from his students, he could actually draw the recognisability map of New York. This is from the original paper. And somehow he showed that well-to-do areas tend to be more recognisable than less well-to-do areas. So this was the first work that was trying to understand the relationship between well-being and let’s say, recognisability of the urban environment.
Now, since then there has not been any work on this quantitative relationship, surprisingly enough. Mainly because it’s difficult to do and lab experiments have their shortcomings. But one of the interesting differences between the ’70s and nowadays is we have the web. We have the internet, right? And as computer scientists, we are actually very happy to use it. And we have the tools to do that and also we could go beyond the lab experiments. And we did a website. It’s called It’s up, the new version right now, if you’re going there. The old version looked like that. So it’s very simple. So we copy Milgram experiment into the web.
Now I’m simplifying that it is not entirely that way, so you have a lot of tricks to do. But it is very simple. You get to show ten pictures from Google street view in London. And you have to guess the closest tube station or borough. Or you can say simply I don’t know. And then based on the correct answers, you can draw a recognisability map of London, as Milgram did, as simple as that. Now what’s the difference? Well, instead of sitting in a classroom for 90 minutes and being subject to the experiment of your professor, its one minute game with a purpose. So you just spend one minute, you share your score on Facebook, on Twitter, and that’s it.
So also the other advantage is that you can stratify the data in all sorts of ways and then present the ways. So in a few months, we had more than 2000 players. And just to give you a really simple example of that, the pink is politically decided boroughs of London. And on the right, you have the cartogram that is distorted, depending on the recognisability scores. And you can clearly see that the central part of London is much more recognisable than other parts. And in terms of collective maps, the south totally disappears.
And as I was telling you, also the cool thing of this is that you can stratify the data by for example, where people are connecting from, from the IP addresses, you can get to a certain extent, their cities. So you know what is recognisable for people inside London, for people outside London, and outside UK. So you can see that recognisability for different five regions of London doesn’t change very much. The ranking doesn’t change. But the location might change. So let me give you a couple of examples. One so I’m going to ask you a couple of questions and you have to be kind enough to guess an answer.
So the first question is so we had location, a specific location that was more recognisable by people outside London than Londoners themselves.
Football stadium. Close enough. Wow. That’s the first time anyone get it. Fantastic. And then the second one was there was at location that was more recognisable by people around the world than Londoners themselves.
Buckingham Palace.
Yeah, but in that case, Londoners would recognise it, right? They might not.
Blackfriars? All right. Well, we had this conversation yesterday. That’s cheating. Right. So as a location this place has been featured in different BBC series, so especially in Sherlock Holmes from the 70s to nowadays. So that’s to tell you that experience of the city is not only creating the physical environment. Sometimes, as computer scientists, when we do experiments, we’re obsessed about doing experiments in situ, right, so being there. But sometimes experience is not always being there. It’s learned through books and films. But so the main point of all this work is that, so there was a relationship between well-being and recognisability. And now for the first time, we actually established that relationship.
So we have enough data to say, OK, we’ve got the recognisability scores of different areas in London and just relate them with the census that you have, in London, UK, well, in the UK, the census data is represented by design D index for multiple deprivation. It tells the extent to which an area is socially deprived or not. And it is a composite index, so it is index divided in different dimensions, income, employment, health, education, housing, living environment, crime. And you can clearly see that there are certain dimensions that are more important, they are more related to recognisability than other dimensions.
So for the first time, you can not only talk about abstract well-being, but you know which dimension of well-being as percentages are related to the recognisability.

What can data from public transport smart cards and a game built on Google Maps tell us about life in the city?

Daniele Quercia shares results from his research, drawing on data from Oyster cards in London, and a game which lets Internet users around the world guess where photos from Google Maps are located, called UrbanOpticon.

Daniele Quercia is a social media researcher at Yahoo Labs in Barcelona. His work has been featured in La Repubblica, The Independent, New Scientist, Le Monde, and on the BBC.

You can watch the whole of Daniele’s presentation “Crowdsourcing for the good of London” on YouTube (21:08).

This article is from the free online

Big Data: Measuring And Predicting Human Behaviour

Created by
FutureLearn - Learning For Life

Our purpose is to transform access to education.

We offer a diverse selection of courses from leading universities and cultural institutions from around the world. These are delivered one step at a time, and are accessible on mobile, tablet and desktop, so you can fit learning around your life.

We believe learning should be an enjoyable, social experience, so our courses offer the opportunity to discuss what you’re learning with others as you go, helping you make fresh discoveries and form new ideas.
You can unlock new opportunities with unlimited access to hundreds of online short courses for a year by subscribing to our Unlimited package. Build your knowledge with top universities and organisations.

Learn more about how FutureLearn is transforming access to education