Rob Kitchin and Tracey Lauriault have just published the second Programmable City Working Paper – Towards Critical Data Studies: Charting and Unpacking Data Assemblages and Their Work. It is a pre-print of a chapter written for the book, Geoweb and Big Data, edited by Joe Eckert, Andy Shears and Jim Thatcher, to be published by University of Nebraska Press.
Abstract
The growth of big data and the development of digital data infrastructures raises numerous questions about the nature of data, how they are being produced, organized, analyzed and employed, and how best to make sense of them and the work they do. Critical data studies endeavours to answer such questions. This paper sets out a vision for critical data studies, building on the initial provocations of Dalton and Thatcher (2014). It is divided into three sections. The first details the recent step change in the production and employment of data and how data and databases are being reconceptualised. The second forwards the notion of a data assemblage that encompasses all of the technological, political, social and economic apparatuses and elements that constitutes and frames the generation, circulation and deployment of data. Drawing on the ideas of Michel Foucault and Ian Hacking it is posited that one way to enact critical data studies is to chart and unpack data assemblages. The third starts to unpack some the ways that data assemblages do work in the world with respect to dataveillance and the erosion of privacy, profiling and social sorting, anticipatory governance, and secondary uses and control creep. The paper concludes by arguing for greater conceptual work and empirical research to underpin and flesh out critical data studies.
Key words
big data, critical data studies, data assemblages, data infrastructures, civil liberties
These seminar videos explore systems such as The City Dashboard and the rise of the Internet of Things (IoT) in terms of data collection, visualization and analysis. Joining these up creates a move towards the Smart City and via innovations in IoT a look towards augmented reality pointing towards the the creation of a ‘Smart Citizen‘, ‘the Quantified Self’ and ultimately a Smart City.
Bio: Dr Andrew Hudson-Smith is Director of the Centre for Advanced Spatial Analysis (CASA) at The Bartlett, University College London, Reader in Digital Urban Systems and Editor-in-Chief of Future Internet Journal. He is also an elected Fellow of the Royal Society of Arts, a member of the Greater London Authority Smart London Board and Course Founder of the MRes in Advanced Spatial Analysis and Visualisation and MSc in Smart Cities at University College London.
Bikeshare initiatives have developed significantly since their introduction in Europe in the 1960’s and are generally regarded as having gone through a number of generations of implementation and design in the interim. The 1st generation, initially deployed in Amsterdam in 1965, was characterized by the use of general purpose bikes, custom painted for identification, and available to the public to borrow from and return to any location. These systems were unmanaged and depended heavily on the integrity of users to use the bikes responsibly. The design ultimately failed as the majority of the fleet was vandalised or stolen.
30 years later Copenhagen introduced the first coin operated or 2nd generation scheme. The construction of the bikes was more robust and the design meant that a degree of control had been introduced. The schemes were expensive however and theft remained a problem due to the fact that users still remained anonymous. In addition, time usage was not limited which meant bikes were frequently kept for extended periods of time making fleet management extremely difficult.
Contemporary, or 3rd generation schemes, have exploited the capacity of information and communications technologies to effectively automate systems and address the shortcomings of previous designs. Networked self-service stations are managed by central computer systems and technologies like RFID are used to monitor the location of bikes, which in turn supports real-time system updates such as the availability of free bikes and stands etc. Users are required to subscribe to the schemes initially and can then access the bikes through a variety of means which include smart cards, fobs, mobile phone applications or even SMS. Knowing who the customer is reduces the likelihood of theft and encourages greater responsibility of the part of the user.
Designs have continued to evolve though and innovations are making way for fourth-generation or demand responsive models which promise not only functional improvements but also the potential to engage with riders in new and more socially progressive ways. While a precise definition of fourth-generation design is still emerging, schemes are characterized by; (a) increased system flexibility; (b) improved distribution; (c) enhanced physical and informational integration with other transportation modes (d) developments such as electric-hybrid bikes and GPS tracking; and (e) increased use of crowdsourcing and participatory platforms.
Susan Shaheen, Director of Innovative Mobility Research at Berkeley, identified BIXI, which launched in Canada in May 2009, as marking the beginning of bikesharing’s fourth generation. A major innovation attributed to the scheme is its use of mobile docking stations which allows stations to be removed and transferred to different locations. This enables stations to be relocated according to usage patterns and user demands. Another feature that could enhance future programs is the use of solar-powered stations. Not surprisingly, solar-powered stations might further reduce emissions and the need to secure access to energy grids to support operations. Fourth generation bikesharing might also consider omitting docking stations altogether which would allow users employ mobile phone technology and street furniture for bicycle pickup and drop-off. These stationless designs are both cheaper and less impactful on the environment while also offering riders higher levels of flexibility and trip customization.
Solar Powered Station at B-Cycle Boulder
Another area of potential improvement for fourth-generation systems is in bicycle redistribution. Geo-fencing for example can directly address this problem, which has dogged the industry for years. By dividing the cycling environment into virtual zones, users can be rewarded when returning bikes to areas that most need them. These zones can be dynamically created or modified as required. The feature is a design element of Socialbicyles (SoBi) in North America.
A third feature of fourth-generation systems is the integration of bikesharing with other forms of transportation. Common payment systems across transit modes enhances usability and reciprocal apps can provide travellers with real-time data on buses, trains, taxis or even carsharing initiatives. This may lead to greater reductions in car ownership as more trips are supported by alternative forms of transport. Creating programs that coordinate meaningful integration is challenging though and often requires multiagency co-operation.
To make bikeshare accessible to more user groups, the use of electric or hybrid bikes may become more widespread in future schemes. Electric bikes overcome what researchers in MIT’s SENSEable City Labidentified as some of the primary obstacles to cycling in an urban environment; distance and topography. Examples of current schemes which have adopted this feature include the recently deployed Bycyken scheme in Copenhagen and BiciMAD in Madrid. And in relation to fleet safety and vandalism, designs which integrate GPS rather than RFID tagging can further deter theft and improve the likelihood of bike recovery. Significantly, active tracking also allows actual routes to be mapped and made available to users. Additionally, it enhances usability by allowing for the possibility of information streams such as; number of miles travelled (per trip and aggregated over time), CO2 offset, calories burned, and money saved vis-à-vis other modes. This information can be a useful way of developing relationships with users and enhancing system usage (e.g. B-Cycles, SoBi).
Socialbicycles – SoBi
The potential of crowdsourcing or participatory sensing may also become a feature of future designs. Individuals could use their personal mobile devices, typically Smartphones, to systematically report on various aspects of their environment. Essentially, the participants themselves either become a spatially distributed sensor network or act to augment sensing technology already incorporated in the bike’s design. Crowdsourcing may also take the form of collaborative map making (Capital Bikes) and route annotation (SoBi) both of which can contribute to system design, infrastructure planning, transportation modelling, and policy formation. The use of social media platforms as a deliberative component of future schemes may strengthen the co-production of services. Using technology this way could create a sense of citizenry where people have meaningful and transformative relationships with the schemes they use.
Exploiting the potential of these innovations to create systems that are integrated, open and participatory will challenge those municipalities which view their role as being primarily concerned with information dissemination and service delivery. This sufficed with traditional 3rd generation schemes which were, for the most part, insular systems designed using quite narrow definitions of efficiency and which offered little possibility for meaningful engagement either with between riders or between riders and decision makers. Next generation schemes will require that civic agencies transform their information and business processes and co-operate with each other in ways that have traditionally been problematic. And for those schemes implemented through partnerships with third parties, public agencies will need to re-think how they select providers, manage contracts, and exploit data. Understanding the response by government agencies to citizen-centric technologies and practices should tell us something about the conditions needed to support social innovation within the sector. It may also tell us how schemes with limited social value are legitimized and how such schemes might be enhanced.
In May 2014 Ubisoft released a new computer game called Watch Dogs. Having sold over 4 million copies in the first week of sales it is tipped to be the game of the year. In the game, Chicago City is controlled by a central operating system (ctOS). The super computer gets a panoptic view of the city using data from cameras and sensor networks. The information obtained is used to manage the city’s infrastructure and technology as well as to maintain a database of personal information about citizens and their activities. In Watch Dogs, a disgruntled computer hacker finds a way to access and hack the ctOS, allowing him to hijack traffic lights, the power grid, bridges and toll gates, rupture water pipes, disable surveillance cameras and access personal information about fellow citizens. The motive for causing mayhem in the city is to find a gang who were involved in his sister’s death and ultimately take down the corrupt system that runs ctOS. In this article, we take a look at some of the real dangers facing today’s cities from malicious hackers.
A Character Accesses City Infrastructure and Data in Watch Dogs
In terms of technology, Chicago, as presented in Watch Dogs is a smart city. Data is fed into the central operating system and the infrastructure of the city adapts and responds accordingly. Although much of the game is fictional, Watch Dogs draws on existing technologies and echoes what is happening today. For example, Rio de Janeiro has a large control centre which applies data analytics to social media, sensors and surveillance cameras in an attempt to predict and control events taking place in the city. Its mission is to provide a safe environment for citizens. Other cities such as Santander and Singapore have invested in sensor networks to record a range of environmental and traffic conditions at locations across the cities. Earlier this year, Intel and Dublin City Council announced that Dublin is also to get a sensor network for measuring city processes. At present many of these projects are focusing on the technical challenge of configuring hardware, designing standards and collecting, storing and processing data from the city-wide sensor networks. Projects are using the data for a range of services to control the city such traffic management (guiding motorist to empty parking spaces), energy management (dimming street lights if no one is present) and water conservation (using sensors to determine when city parks need water).
The Internet of Things & Security
The roll out of such smart city technology is enabled through the Internet of Things (IoT) which is essentially a network of objects which communicate and transfer data without requiring human-to-human or human-to-computer interaction. The IoT can range from a pace maker sending patient information to a physician, to a fridge informing its owner that the milk is low. In the smart city, sensors automatically relay data to a control centre where it is analysed and acted upon.
The Control Centre in Rio de Janeiro
While Watch Dogs raises important moral and ethical issues concerning privacy and the development of a big brother society via smart city technologies, it also raises some interesting questions about the security and reliability of the technology. In Watch Dogs, the main character can control the city infrastructure using a smart phone due to a security weakness in the ctOS. In reality, we have recently seen objects in the IoT being compromised due to weaknesses in the hardware security. Baby monitoring webcams which were accessed by hackers and demonstrations of how insulin pumps can be compromised are cases which have received media attention. Major vulnerabilities of the IoT, were seen in late 2013 and early 2014 when an orchestrated cyber attack saw 100,000 ‘things’ connected to the Internet hacked and used to send malicious spam emails. The hacked ‘things’ included smart TVs, fridges and media centres. Basic security misconfigurations and failures to alter default passwords left devices open to attack.
Even mature internet technologies such as those used in ecommerce websites are vulnerable to hacking. In May this year e-bay’s web servers were hacked leading to the loss of user data. Security flaws with the OpenSSL cryptography standard (used to transmit data securely on the Internet) came to light in April 2014 with the ‘Heartbleed’ bug. A vulnerability enabled hackers to access the short term memory of servers to capture information such as passwords or credit card details of users who recently interacted with the server. All technologies which can send and receive data are vulnerable to attack and misuse unless strict security protocols are used and kept up-to-date. Unfortunately, as the examples here highlight, it seems that the solutions to security issues are only provided after a problem or a breech has been detected. This is because it’s often an unknown bug in the code or poor coding practice which provides a way for hackers to access systems remotely. There is a reluctance to invest in thorough testing of technologies for such weaknesses. Development companies seem prepared to risk cyber attacks rather than invest in the resources required to identify problem areas.
Hacking the Smart City
The fact that all systems connected to the Internet appear vulnerable to cyber attacks is very worrying when considered in the context of smart cities. Unlike personal webcams, TVs and fridges, the technology of smart cities forms part of a complex and critical infrastructure which is used to calibrate and control a city.While governments and city authorities are generally slow to publicise attacks on their technological infrastructure, the Israeli government has acknowledged that essential services that run off sensors, such as water, electricity and banking, have been the target of numerous hacking attacks. For example, in 2013, the traffic management system for a main artery in the port city of Haifa, was hacked, causing major traffic problems that lasted for several hours. Such malicious hijacking of technology is inconvenient for citizens, costs the city financially and could also have fatal consequences. Last year, it was demonstrated that it was relatively easy to hack the traffic light system in New York City. By sending false signals regarding the traffic flow at particular junctions, the algorithm used to control the traffic light sequence could be outsmarted and fooled into thinking that a particular junction was busy and therefore adjust the green time of traffic lights in a particular direction.
City technology is built on legacy systems which have been incrementally updated as technology has changed. Security was often not considered in the original design and only added after. This makes such legacy systems more vulnerable for exploiting. For example, many of the traffic light coordination systems in cities date from the 1980s when the main security threat was physical interference. Another problem with the city technology is the underlying algorithms which can be purely reactive to the data they receive. If false data is supplied then the algorithm may produce undesirable consequences. While the discussion here has focused on sensors embedded in the city, other sources of data, such as social media are open to the same abuse. In March 2014, the twitter account of The Associated Press was hacked and a message reporting of an attack on President Barrack Obama was posted. This led to $136 billion being wiped of the NY stock exchange within seconds. This is an example of humans using bad data to make a bad decision. If the human cognition process is unable to interpret bad data, what hope do pre-programmed computer algorithms have?
As cities continue to roll out technologies aimed at enhancing the lives of citizens, they are moving towards data driven forms of governance both for long term and short term actions. Whatever type of sensor is collecting data, there is a danger that data can be biased, corrupt, played, contained errors or even be faked through hacking. It is therefore imperative for city officials to question the trustworthiness of data used in decision making. From a technical point of view, the data can be made safe by calibrating the sensors regularly and validating their readings against other sensors. From a security perspective, the hardware needs to be secured, maintained and updated to prevent malicious hacking of the device. Recognising the threat which has been highlighted by Watch Dogs, the US Centre for Internet Security (CIS) issued a Cyber Alertregarding the game stating that ‘CIS believes it is likely that a small percentage of Watch Dog players will experiment with compromising computers and electronic systems outside of game play, and this activity will likely affect SLTT (State, Local, Tribal and Territorial) government systems and Department of Transportation (DOT) systems in particular.’
In other domains, such as the motor industry there is a move to transfer functions from the human operator to algorithms. For example, automatic braking, parking assistance, distance based cruise control and pedestrian detection are becoming mainstream in-car technologies in a slow move towards vehicles which drive themselves. It is likely that managing the city will follow the same pattern and incrementally the city will ‘drive’ itself and could ultimately be completely controlled by data-driven algorithms which react to a network of sensors. Although agencies such as the CIS give some advice to minimise the risk of Cyber Attacks on cities, it seems that hacking of the smart city infrastructure is inevitable. The reliance of cities on software and the risks associated with this strategy are well known (Dodge & Kitchin, 2004; Kitchin, 2014). The problem is compounded by the disappearance of the analogue alternative to smart city technologies (Townsend, 2013). This could lead to prolonged recovery from attacks and bugs due to the total reliance on technology to run cities. Cities therefore need to consider the security risks connected to deploying and using sensors to control a city and make decisions. It is also essential that control loops and contingency plans are in place to allow a city to function during a data outage just as contingency plans are made for handling the loss of other essential services such as power and water.
References
Dodge, M., & Kitchin, R. (2004). Flying through code/space: The real virtuality of air travel. Environment and Planning A, 36(2), 195–211.
Townsend, A. (2013). Smart cities: Big data, civic hackers, and the quest for a new utopia. New York: W.W. Norton & Co.
This video introduces the Programmable City. In the video Rob Kitchin outlines the aims and objectives of the project and highlights the importance of the support received from the European Research Council (ERC). Each researcher on the Programmable City team also briefly discusses their work. The video was made by the Programmable City team in order to promote the project and the ERC.
