Friday 20 April 2012, Senate House, University of London
By Koen van Dam
CASA, the Centre for Advanced Spatial Analysis at UCL chose “Smart Cities: bridging physical and digital” [pdf] as the theme of their conference held at the Senate House in Central London on the 20th of April 2012. Smart cities and digital… of course DCE had to be there.
Prof Michael Batty, chair of the management board of CASA, opened the conference by going back 40 years in the past to highlight the many advances of the digital age, most notably the rise of the internet, and how they have changed life in cities. Batty went on to explain that cities can be considered as networks of connected computers, and that smart cities present planners with new challenges because they address the short term operation of the places we live in rather than long-term strategies. The question is now what the next 40 years will hold. The presentations at this conference might offer a glimpse of that future. The real challenge, according to Batty, is that after the transition from “real” to “digital”, we now have to move back from “digital” to “real” and to see the changes of digital services on daily life.
The first speaker of the conference was Prof Carlo Ratti, Director of MIT’s SENSEable City Lab. He provided an image which would turn out to be a metaphor central to the whole day: the idea that a F1 racing team can no longer just win races by having the best car and the best driver, but that real-time data processing of all things happening in the car and on the track is absolutely essential now. Bringing this back to cities: they have become control systems (also strengthening Batty’s claim about short term operation being more central). Prof Ratti spent most of his talk introducing wonderful examples of analysis of data, digital design and real applications, ranging from analysing telecom signals during the world cup football, measuring drought by counting green pixels in digital photos, interactive walls made of water to early prototypes of the Copenhagen Wheel, a way to store power in a bicycle while also providing the rider with real-time data on for example air quality. Finally, Ratti made very clear that in the past key publications were made by one author from one discipline, but that today the most influential articles are in fact written by many authors coming from various scientific backgrounds.
The other speakers of the day were all researchers at CASA. Jon Reades and Joan Serras gave a great overview of the analysis of transport data (e.g. Oyster journeys on London’s transport network) with many beatiful maps. James Cheshire and Martin Austwick looked at cycle hire schemes, showing how much we can learn and infer from just looking at the status of docking stations around the city without having to know exactly which trips were made. One of the key advantages of using OpenStreetMap (instead of for example Google maps) is that you get a lot more information about the usability of the roads for, for example, cyclists, and the team took this into account in their routing algorithms.
Next up was a slightly controversial, but highly interesting talk by Prof Sir Alan Wilson, who developed a model of the riots in London August 2011. Using epidemic models (c.f. Epstein) they try to replicate the attractiveness of certain site to rioters and looters, taking for example the number of police officers present into account to determine the chance for arrest. One of the members of the audience, quite rightfully, pointed out that in some riots a high police presence is actually the cause for escalation instead of serving as a way to keep people under control. Also, the highly political impact of research would make it hard to develop objective simulation models. Still, this was one of the few examples at the conference of using data to inspire predictive models and using them for decision support.
George MacKerron showed us Mappiness, a way to study how people feel in certain locations based on an iPhone app which asks people to state their own happiness after having been prompted at random times. Analysis of this data could show in which environment people feel most happy.
Richard Milton discussed data stores and real-time data, raising the all-important question of data analysis: how to find something you don’t know was there. Organising data by spatial elements, it becomes, for example, possible to overlay population density maps with energy infrastructures for gas and electricity, showing striking matches. Maptube a tool developed at CASA for mapping data sets and building mash-ups, was introduced after which Steven Gray took over the microphone to present another tool offered by CASA, namely GEMMA engine for mapping. The idea is that these tools could become a “Big Data Toolkit” for visualisation and analysis of data. As a perfect illustration of the kind of events we might be able to infer from data, a lady on the tube wondering “it’s not usually this busy, what happened??” made clear that we need predictions as well as data analysis to get a better overview of systems.
Andy Hudson-Smith, CASA director, showed the Tales of Things project, explained as the “Internet of Second-hand Things”, paraphrasing the concept of the Internet of Things. Oxfam used this approach in their charity shops, allowing people to tell stories about the objects they were donating.
The final speaker at the conference was mapping guru Oliver O’Brien. He showed the CityDashboard displaying in real-time a number of views of a city, enabling the user to discover if there is something “wrong” (e.g. by linking travel disturbances, weather and popular news items) and help to make the right decisions on for example departure time or routing. Perhaps a bit disappointingly, the dashboard does not contain any “intelligence” or predictions, but the presentation gave a valuable overview of the data used as well as the standards in which they are made available, and the API would enable others to build on top of this dashboard by accessing the same information. Furthermore, Oliver highlighted that in order to compare different cities, more standards are required.
During lunch and tea breaks, participants were able to see the brilliant visualisations, apps and games developed by the CASA researchers up close and personal. Many of the stands were interactive, and we flapped our arms as we flew over London, we watched air planes circle over the city, placed police units in riot hot spots, looked at agent-based models of people moving about, and studied how pedestrians changed their behaviour as we adjusted the urban footprint of their virtual world using wooden blocks. Furthermore, it was a chance to stare at the recently released map of every bus trip in London and several other impressive visualisations at the border of art and science.
The conference ended with a panel discussion, in which several of the speakers took place to discuss issues such as which real insights can we get from visualisation and which sectors they predict would be the next big application after transport (which was clearly the most studied subject). In answering these questions, the panel members discussed the gap between sensing to activating and put forward health and justice (Alan Wilson), education (Ratti) and retail (Batty) as future domains to study.
On a final note, two statements of great importance to our Digital City Exchange programme were made during the panel discussion. Batty stressed that while we have a lot of data, we still know very little about the processes underlying decisions and actions of people. Hudson-Smith then said that what is needed next is an integrated model bringing together things that happen in different sectors. We are working on it!