The real SimCity
‘This is the Notting Hill Carnival,’ says Professor Mike Batty. But this isn’t the West London street party with which most people are familiar – there are no high-volume sound systems, no food stalls selling jerk chicken and no parades of floats overloaded with sequined dancers, just hundreds of little red dots rushing excitedly about on a computer screen.
This computer simulation of the carnival was created by Batty, head of London’s Centre for Advanced Spatial Analysis (CASA), in an attempt to solve the overcrowding problem at one of the world’s biggest street parties. Each red dot represents a person and, unlike in real life, the movements of these virtual partygoers can be manipulated with just a few taps of a computer key.
‘If we increase the number of people coming into an entrance, we can see how that would change the crowding points,’ Batty explains as he unleashes a herd of virtual partygoers and manipulates where they go by opening up virtual entrances, closing down virtual streets and even changing the parade route completely. ‘It’s all about asking, “What if we do this?”, or, “What if we do that?”,’ he says.
Model behaviour
Batty is one of a handful of British digital geographers who are creating computer simulations of real-life geographical areas in order to solve real-life problems. The computers analyse data to create rules about how humans or other phenomena behave and then use these rules to predict the future by asking ‘what if’ questions. Get the answers right – and that’s no easy task – and it should enable policy-makers and planners to test decisions virtually in order to avoid making potentially costly mistakes in the real world.
The range of questions they can attempt to answer is huge. Which parts of a London will be flooded if the oceans rise by one metre? What will happen to pollution levels if the congestion charge is removed? Where is an immigrant who enters a city today likely to be living in ten years’ time? What type of person is most likely to move into the new flats that are being built opposite my house?
The data can be taken from all sorts of places. The Notting Hill Carnival model was pulled together during 2000 from four sources: a street survey, tube station entrance/exit counts, St John Ambulance and a police helicopter. ‘We had a limited amount of data so it produced a limited model,’ Batty explains. ‘But nobody would have thought of doing this ten years ago. We just didn’t have the data or the computers.’
At Leeds University, Dr Mark Birkin is attempting to build a hugely ambitious model of the whole of Britain from the 2001 census information, while also creating demonstrators to solve specific problems in the city where he works. ‘Leeds City Council is suffering from a data deluge; there are data everywhere about the people and the city,’ he says. ’Using models, we can make sense of that data and pick out what’s important.’
Stephen Boyle, chief regeneration officer for Leeds City Council, hopes to find out who will end up living in the planned new developments as part of a scheme to regenerate one tenth of the city. Leeds has a large immigrant population and increasing support for the British National Party in certain areas, so it’s vital he can attract a mixed population to these developments in order to ensure that the new communities are successful. ‘If you have a mix of different incomes, different house tenures, different cultures, you create a more sustainable area,’ he explains.
‘We’re talking to Dr Birkin because we’re interested in finding out who will come into these developments and what impact that is likely to have on the area,’ Boyle continues.
The models are already predicting city-wide trends – a growing immigrant community and an increasing elderly population – but Boyle hopes they can be used to solve more specific problems. ‘At the moment, the jury’s out. I’m sure it will be useful, but it could be very useful.’
More than a game
People are comparing these new models to the computer game SimCity, in which players run an onscreen city and wait to see if their policy decisions cause their settlement to prosper or collapse. But Birkin says there are big differences between the computer game and his own models. ‘You can’t push a button and win or lose. I don’t think real life is like that. It’s about evaluation of options and trying to make robust decisions in uncertain circumstances.’
What the models are already revealing is the true nature of cities. What was once seen as a chaotic sprawl is now being re-evaluated as misunderstood complexity. ‘The messiness isn’t as messy as we might think. There’s a logic to it,’ says Batty. ‘Cities are the result of lots of people making millions of decisions, and there’s been a shift from thinking about a city as an organism controlled from the top to something that is built from the bottom up by all of these relatively uncoordinated decisions.’
Models can pick out the patterns in these decisions and help planners work with the resident population’s behaviours rather than against them. Back on screen, maps generated by Batty’s models show how far people in London currently travel to work and, with a few changes, how this would be affected if public transport costs were lowered or if a new development was constructed that provided local jobs.
Batty hopes these models will ‘stop us doing stupid things’ and show ‘where we need to work harder at economic development’. He points to the government’s plans to build in the Thames Gateway, despite the fact that most work is to the west and, he says, ‘the DLR [Docklands Light Railway] is a joke’. ‘These models can tell us what kind of subsidies we might need to generate jobs and people in locations that are currently unattractive,’ he says.
And he hopes that they will also rein in some of the grander fantasies of architects and developers – such as the self-contained eco-cities planned for China. ‘This Dongtan eco-city thing in China is a pie in the sky,’ he says. ‘The models don’t take account of how people live and work. Everywhere in London, people are travelling across the city to work, but plans for these eco-cities say people will live and work in the same area. How are they going to make people do that?’
London at your fingertips
The number-crunching is just one side of the future-predicting models being created by British geographers. Their results need to be communicated, and ‘when you have very complex things, the best way to understand them is to visualise them,’ says Batty. ‘So we’re very interested in trying to visualise them in both two and three dimensions.’
Step forward Virtual London: a 3D map of Britain’s capital created by a team led by Dr Andrew Hudson-Smith, a senior research fellow also working at CASA with Batty. Looking at it on screen is just like looking at a computer game. It contains a 3D representation of every single building in London, geographically located in the correct place, stretching right out to the M25.
Virtual visitors can fly over the rooftops of the city, duck down over the surface of the River Thames and then zoom up over the Houses of Parliament. Or should they wish, they can ‘explode’ buildings such as the ‘Gherkin’ tower and watch them tumble in on themselves until there’s nothing left except a gaping hole in the city skyline.
It’s entertaining, but it was created with a more serious purpose in mind: to test a myriad of ‘what if’ scenarios regarding planning in the city of London. Users can virtually build or destroy any building in London with just a few keystrokes. The plan was to put it online and get the general public more involved in shaping the architectural future of the city in which they live.
‘Our thoughts were that if you look around London, there are these 1960s and ’70s cock-ups that should never have happened,’ explains Hudson-Smith. He believes this urban acne could have been avoided if the public were more informed about development plans. ‘Yes, there’s always some note stuck on a lamppost somewhere, saying a planning application has been sent in, and you can go and look at it at the local town hall, but nobody ever does.’
And even if they do, they will just see the architect’s vision of what they hope a development will look like. ‘It’s all marketing,’ says Hudson-Smith. ‘In front of a planned supermarket there are always wealthy-looking people; the sky is always blue; there are hot-air balloons floating around. We wanted to put something out there that showed London as it is. So it’s a bit grey. And it’s beautiful, but it’s also a bit crap in places.’
The original model was funded by the former mayor of London, Ken Livingstone, with the intention of changing the planning law so that every application would have to be submitted with a virtual plan. These could then be projected onto the virtual cityscape to be judged and commented on by the public. ‘It puts the public in charge – and the public should have a free say because we live here,’ says Hudson-Smith.
Unfortunately, after six years of research, the plans have had to be put on hold because of issues with the licensing system of the Ordnance Survey, whose maps Virtual London uses as a base. ‘It’s a ridiculously arcane system,’ says Hudson-Smith, whose team is now working on a model of Phuket in Thailand that doesn’t suffer from such licence restrictions.
More data, more problems
For all of the digital geographers creating predictive computer models, this is really just the beginning. ‘The main sticking point is the translation from academic into practice. We just don’t train enough people in it currently,’ says Batty. ‘There are so many skills involved. It isn’t just data – we need mathematicians, social scientists, geographers…
‘Worldwide, the number of people working on these subjects in increasing,’ he continues. ‘We’re getting more data and producing better models. But the problems we’re trying to solve are getting more severe at the same time. It’s a process of co-evolution. We’re always on the edge of what we can predict and we always want to be able to predict more.’
Finding our way around 3D worlds
Just as people have had to learn to use 2D maps, they’re now having to learn to use 3D representations of our world.
Dr Andrew Hudson-Smith encountered problems when people viewing his 3D London air pollution map, where higher levels of pollution were indicated with higher 3D peaks. ‘Some people thought that all they needed to do to escape the pollution was fly up higher where the air was clear, and you think, “Hmmm. No, that’s not how it works”,’ he says.
‘It’s quite difficult to get people to walk through a visualisation at the pace people really walk through a city because it just seems too slow,’ says Dr Jennifer Whyte, author of Virtual Reality and the Built Environment. She believes this is because many of us have become used to whizzing around computer games.
This familiarity with computer landscapes can lead to further disappointments with 3D virtual worlds based on reality. ‘If you watch Lara Croft going through one of these games, she’s in cavernous spaces,’ says Whyte. Gaming landscapes are scaled up to compensate for small screens and a lack of peripheral vision, she explains. ‘In comparison, if you have a real city at a real scale, it may seem quite pokey.’
August 2008
This computer simulation of the carnival was created by Batty, head of London’s Centre for Advanced Spatial Analysis (CASA), in an attempt to solve the overcrowding problem at one of the world’s biggest street parties. Each red dot represents a person and, unlike in real life, the movements of these virtual partygoers can be manipulated with just a few taps of a computer key.
‘If we increase the number of people coming into an entrance, we can see how that would change the crowding points,’ Batty explains as he unleashes a herd of virtual partygoers and manipulates where they go by opening up virtual entrances, closing down virtual streets and even changing the parade route completely. ‘It’s all about asking, “What if we do this?”, or, “What if we do that?”,’ he says.
Model behaviour
Batty is one of a handful of British digital geographers who are creating computer simulations of real-life geographical areas in order to solve real-life problems. The computers analyse data to create rules about how humans or other phenomena behave and then use these rules to predict the future by asking ‘what if’ questions. Get the answers right – and that’s no easy task – and it should enable policy-makers and planners to test decisions virtually in order to avoid making potentially costly mistakes in the real world.
The range of questions they can attempt to answer is huge. Which parts of a London will be flooded if the oceans rise by one metre? What will happen to pollution levels if the congestion charge is removed? Where is an immigrant who enters a city today likely to be living in ten years’ time? What type of person is most likely to move into the new flats that are being built opposite my house?
The data can be taken from all sorts of places. The Notting Hill Carnival model was pulled together during 2000 from four sources: a street survey, tube station entrance/exit counts, St John Ambulance and a police helicopter. ‘We had a limited amount of data so it produced a limited model,’ Batty explains. ‘But nobody would have thought of doing this ten years ago. We just didn’t have the data or the computers.’
At Leeds University, Dr Mark Birkin is attempting to build a hugely ambitious model of the whole of Britain from the 2001 census information, while also creating demonstrators to solve specific problems in the city where he works. ‘Leeds City Council is suffering from a data deluge; there are data everywhere about the people and the city,’ he says. ’Using models, we can make sense of that data and pick out what’s important.’
Stephen Boyle, chief regeneration officer for Leeds City Council, hopes to find out who will end up living in the planned new developments as part of a scheme to regenerate one tenth of the city. Leeds has a large immigrant population and increasing support for the British National Party in certain areas, so it’s vital he can attract a mixed population to these developments in order to ensure that the new communities are successful. ‘If you have a mix of different incomes, different house tenures, different cultures, you create a more sustainable area,’ he explains.
‘We’re talking to Dr Birkin because we’re interested in finding out who will come into these developments and what impact that is likely to have on the area,’ Boyle continues.
The models are already predicting city-wide trends – a growing immigrant community and an increasing elderly population – but Boyle hopes they can be used to solve more specific problems. ‘At the moment, the jury’s out. I’m sure it will be useful, but it could be very useful.’
Start the slideshow (5 pictures)
More than a game
People are comparing these new models to the computer game SimCity, in which players run an onscreen city and wait to see if their policy decisions cause their settlement to prosper or collapse. But Birkin says there are big differences between the computer game and his own models. ‘You can’t push a button and win or lose. I don’t think real life is like that. It’s about evaluation of options and trying to make robust decisions in uncertain circumstances.’
What the models are already revealing is the true nature of cities. What was once seen as a chaotic sprawl is now being re-evaluated as misunderstood complexity. ‘The messiness isn’t as messy as we might think. There’s a logic to it,’ says Batty. ‘Cities are the result of lots of people making millions of decisions, and there’s been a shift from thinking about a city as an organism controlled from the top to something that is built from the bottom up by all of these relatively uncoordinated decisions.’
Models can pick out the patterns in these decisions and help planners work with the resident population’s behaviours rather than against them. Back on screen, maps generated by Batty’s models show how far people in London currently travel to work and, with a few changes, how this would be affected if public transport costs were lowered or if a new development was constructed that provided local jobs.
Batty hopes these models will ‘stop us doing stupid things’ and show ‘where we need to work harder at economic development’. He points to the government’s plans to build in the Thames Gateway, despite the fact that most work is to the west and, he says, ‘the DLR [Docklands Light Railway] is a joke’. ‘These models can tell us what kind of subsidies we might need to generate jobs and people in locations that are currently unattractive,’ he says.
And he hopes that they will also rein in some of the grander fantasies of architects and developers – such as the self-contained eco-cities planned for China. ‘This Dongtan eco-city thing in China is a pie in the sky,’ he says. ‘The models don’t take account of how people live and work. Everywhere in London, people are travelling across the city to work, but plans for these eco-cities say people will live and work in the same area. How are they going to make people do that?’
London at your fingertips
The number-crunching is just one side of the future-predicting models being created by British geographers. Their results need to be communicated, and ‘when you have very complex things, the best way to understand them is to visualise them,’ says Batty. ‘So we’re very interested in trying to visualise them in both two and three dimensions.’
Step forward Virtual London: a 3D map of Britain’s capital created by a team led by Dr Andrew Hudson-Smith, a senior research fellow also working at CASA with Batty. Looking at it on screen is just like looking at a computer game. It contains a 3D representation of every single building in London, geographically located in the correct place, stretching right out to the M25.
Virtual visitors can fly over the rooftops of the city, duck down over the surface of the River Thames and then zoom up over the Houses of Parliament. Or should they wish, they can ‘explode’ buildings such as the ‘Gherkin’ tower and watch them tumble in on themselves until there’s nothing left except a gaping hole in the city skyline.
It’s entertaining, but it was created with a more serious purpose in mind: to test a myriad of ‘what if’ scenarios regarding planning in the city of London. Users can virtually build or destroy any building in London with just a few keystrokes. The plan was to put it online and get the general public more involved in shaping the architectural future of the city in which they live.
‘Our thoughts were that if you look around London, there are these 1960s and ’70s cock-ups that should never have happened,’ explains Hudson-Smith. He believes this urban acne could have been avoided if the public were more informed about development plans. ‘Yes, there’s always some note stuck on a lamppost somewhere, saying a planning application has been sent in, and you can go and look at it at the local town hall, but nobody ever does.’
And even if they do, they will just see the architect’s vision of what they hope a development will look like. ‘It’s all marketing,’ says Hudson-Smith. ‘In front of a planned supermarket there are always wealthy-looking people; the sky is always blue; there are hot-air balloons floating around. We wanted to put something out there that showed London as it is. So it’s a bit grey. And it’s beautiful, but it’s also a bit crap in places.’
The original model was funded by the former mayor of London, Ken Livingstone, with the intention of changing the planning law so that every application would have to be submitted with a virtual plan. These could then be projected onto the virtual cityscape to be judged and commented on by the public. ‘It puts the public in charge – and the public should have a free say because we live here,’ says Hudson-Smith.
Unfortunately, after six years of research, the plans have had to be put on hold because of issues with the licensing system of the Ordnance Survey, whose maps Virtual London uses as a base. ‘It’s a ridiculously arcane system,’ says Hudson-Smith, whose team is now working on a model of Phuket in Thailand that doesn’t suffer from such licence restrictions.
More data, more problems
For all of the digital geographers creating predictive computer models, this is really just the beginning. ‘The main sticking point is the translation from academic into practice. We just don’t train enough people in it currently,’ says Batty. ‘There are so many skills involved. It isn’t just data – we need mathematicians, social scientists, geographers…
‘Worldwide, the number of people working on these subjects in increasing,’ he continues. ‘We’re getting more data and producing better models. But the problems we’re trying to solve are getting more severe at the same time. It’s a process of co-evolution. We’re always on the edge of what we can predict and we always want to be able to predict more.’
Finding our way around 3D worlds
Just as people have had to learn to use 2D maps, they’re now having to learn to use 3D representations of our world.
Dr Andrew Hudson-Smith encountered problems when people viewing his 3D London air pollution map, where higher levels of pollution were indicated with higher 3D peaks. ‘Some people thought that all they needed to do to escape the pollution was fly up higher where the air was clear, and you think, “Hmmm. No, that’s not how it works”,’ he says.
‘It’s quite difficult to get people to walk through a visualisation at the pace people really walk through a city because it just seems too slow,’ says Dr Jennifer Whyte, author of Virtual Reality and the Built Environment. She believes this is because many of us have become used to whizzing around computer games.
This familiarity with computer landscapes can lead to further disappointments with 3D virtual worlds based on reality. ‘If you watch Lara Croft going through one of these games, she’s in cavernous spaces,’ says Whyte. Gaming landscapes are scaled up to compensate for small screens and a lack of peripheral vision, she explains. ‘In comparison, if you have a real city at a real scale, it may seem quite pokey.’
August 2008
|
FIND OUT WHAT WE'RE TALKING ABOUT ON TWITTER:
| |
 |
