This essay was originally published at Next City. I’m especially grateful to Diana Lind for offering the space to reach a new readership, and to Ariella Cohen for her excellent edits, which made my quirky style of writing much more comprehensible to a general audience. I’m publishing the original version of the essay here in order to preserve a copy on my own server as well as to preserve the multimedia for all you fellow fans of The Wire.
Self Knowledge Through Numbers
Over the past few years a merry band of geeks from around the world has given rise to the movement of the quantified self. They explain their mission as “self knowledge through numbers,” though Vanity Fair sarcastically calls them “weirder, hive minder weight watchers.”
The basic premise of the quantified self is perhaps best summed up by a popular slogan from business consultant Peter Drucker: “What gets measured gets managed.” If we aspire to run faster, then we must use a stopwatch to time our pace. If we want to lose weight, then we must buy a scale to measure our progress and adapt our diet until we reach our goal. Modern self-trackers also measure their sleep, meals, meditation, heart rate, moods, finance, blood pressure, and even amino acids.
The rise of the quantified self is indicative of our more general obsession with data; but it also illustrates our growing concern with personal perfection at the expense of our community’s well being. Popular books — like Timothy Ferriss’ 4-Hour Workweek — advise readers to shut out the distractions of the wider world in order to focus entirely on self-improvement and personal perfection. If what gets measured gets managed, then, increasingly, we are managing ourselves without considering the needs of our neighbors and communities.
The Quantified City
What if we were to apply the model of the quantified self to the development of our cities? It’s a question that appears to be gaining steam. Esther Dyson, an influential angel investor and technology analyst, has observed the emergence of a suite of applications that enable citizens and governments to monitor the ‘health’ of their communities. Civic Insight, for example, has partnered with New Orleans to enable citizens to monitor what the local government is doing to address the city’s blight. Yelp has partnered with New York and San Francisco to make restaurant inspection data available on restaurant profile pages. (Boston, Philadelphia and Chicago have already committed to making their restaurant inspection data available using the same standard.) The Daily Brief allows residents of Baltimore, Bloomington and Boston to monitor all the 311 service requests made by citizens each day. GreatSchools is to schools what Yelp is to restaurants; it offers parents comprehensive profiles of local schools, including detailed information about test scores, enrollment statistics, student demographics, and annual budgets. Each school profile on GreatSchools even incorporates data from Zillow so that parents can search for available homes in the best school districts.
These platforms can empower citizens in two ways. First, they are able to make better decisions about where they send their children to school, where they choose to live, how they get to work, and how to stay healthy. (This goes beyond avoiding restaurants with poor inspection scores; researchers have also found a correlation between Yelp ratings for hospitals and lower patient readmission and mortality rates.) Second, such platforms are tools that empower citizens and civil society organizations to monitor the performance and spending of public services to demand for greater accountability.
The smart city narrative has been defined by commercial actors like Cisco, Siemens and IBM that set out to enchant politicians with visions of a fully automated, sensor-enabled urban utopia. The two most illustrative examples are Masdar City, a walled suburban enclave outside Abu Dhabi, and Songdo, a new South Korean city 35 miles south of Seoul that is slated to open in 2015. These cities are smart indeed. Streetlights only turn on when they detect movement. Escalators are activated when they detect someone approaching. Homes are wired for “telepresence videoconference systems” that connect to local hospitals, schools, and government offices. Trash cans measure daily waste and automatically signal when collection vehicles should pass by.
This is only the beginning. In a few years time we should expect our home thermostats — such as the Nest, developed by former Apple engineers — to share our energy consumption not just with our energy utility, but also with our neighbors. Our garden sprinklers will connect to the weather report to take into account tomorrow’s rain storm before soaking our plants today. Our homes and offices will effectively ‘hibernate’ once the last family member or employee is out the door, and local governments might give us tax breaks for taking shorter showers or leaving the refrigerator door open less often — all of which will likely be detected by pervasive, connected sensors.
We all want to live in the future; just as we all fantasize about escaping to the past. But beyond the “oh neat” factor, what’s so great about living in a city where everything is connected to everything else? Will our lives actually improve?
An increasing number of social critics and urban planners are speaking out against the smart city. “No one likes a city that’s too smart,” notes the sociologist Richard Sennett in an essay that lambasts the development model of Masdar City and Songdo. Writing in the Boston Globe, author Courtney Humphries has collected the various arguments of critics of the “too-smart city.” They remind us that we have already tried to build smart cities from scratch using the latest technologies of the past, and that those failures should serve as warning signs from the past as cities invest billions of dollars in top-down projects based on the technologies of today.
During the 1950s — at opposite ends of the earth — two architects built massive cities from scratch: Oscar Niemeyer’s Brasilia, the capital of Brazil, and Le Corbusier’s Chandigarh, a symbol of newly independent, modern India. These two new cities — like the post-war suburbs that mushroomed throughout North America and Western Europe — were built around the new technology of the day, the automobile. Today both cities are model case studies in failed urban planning. Brasilia is among the world’s least pedestrian-friendly cities, and Chandigarh has more cars per person than any other Indian city.
There are few sidewalks or pedestrian bridges in Brasilia, but well worn footpaths are observable from satellite photography. They reveal the complicated paths that pedestrians must take just to cross one of the city’s many busy arteries.
The automobile of the 1950s was the symbol of freedom, modernity and convenience. But neither Niemeyer nor Le Corbusier nor the majority of urban planners of their generation foresaw that car-centered city planning would result in gridlock, pollution, road rage, drunk driving, and social isolation. The failures of car-centered urban planning should serve as notes of caution for cities that want to develop around specific technologies.
The sensor-based city leads us to think less, not more. In certain circumstances, this strikes me as a reasonable proposition. I don’t want to invest my limited cognitive energy in searching for parking or reducing my energy consumption. If technology is able to lend a helping hand, then I gladly accept. However, there is a real risk that as our lives become more automated we become more like automatons. We will follow the guidance of our smart phones without reflecting on how we live our lives, and how we engage in our communities. Furthermore, there is a risk that the smart city draws attention and resources to those issues that can be easily quantified and calculated — for example, how to most efficiently re-route traffic — at the expense of complex issues that aren’t easily quantified — such as how to create public spaces that promote social inclusion.
Smart cities could lead to dumb citizens.
An alternative approach is to employ new technologies that prioritize smart citizens over smart cities. In her critique of the “too-smart city,” Humphries points to Boston’s Office of New Urban Mechanics as a “deliberate counterweight to the heavy-handed, control-room vision of urban technology.” New Urban Mechanics, which launched a sister project in Philadelphia, focuses on “participatory urbanism” as a way to involve citizens in the planning of their city and improvement of their communities. Similar initiatives have launched in San Francisco — where citizens help the city government to improve access to healthy food or reduce its environmental footprint — and New York where a grassroots group of residents re-imagine blighted public spaces as opportunities for so-called “placemaking.”
A pedestrian underpass as it currently stands in Crown Heights, Brooklyn, New York
The same pedestrian underpass as re-imagined by a group of participatory urbanists.
The Latin American Network for Sustainable, Democratic and Just Cities offers another model to involve citizens in the improvement of their communities. Over 70 different citizen observatories belong to the network. They include Jalisco Como Vamos (“Jalisco, how are we doing?”) in Mexico, Nossa São Paulo in Brazil, Nuestra Buenos Aires in Argentina, and dozens more throughout the region.
Often tied to a local university, these citizen observatories track 100 basic indicators from official government data sources. The indicators range from the average percentage of income spent on housing to the amount of green space per resident to statistics about employment, pollution, commute times, health, education, equality and public security. In addition to information from official government sources, such as census data, the citizen observatories also contract with a polling company to survey a representative sample of the city’s population to document their perceptions of urban quality of life. This methodology emphasizes the importance of citizen perception. It’s not enough for a city’s homicide rate to fall; citizens need to feel safer as well. In Brazil, over 500 mayoral candidates signed agreements to work hand-in-hand with these citizen observatories when drafting their city plans and monitoring their implementation.
An infographic from the annual report of Jalisco Como Vamos describes a number of indicators related to housing in Mexico’s Jalisco State. Of the 1.8 million homes and apartments in the state, more than 350,000 are uninhabited. More than 50% of homes are constructed informally without the required permits.
The Global City Indicators Facility is a similar network of 250 cities worldwide that submit dozens of indicators each year. Unlike the Latin American Network for Sustainable, Democratic and Just Cities, however, the exercise doesn’t inform civic participation or advocacy initiatives.
There is nothing wrong with installing sensors within trash cans, parking spaces, and sewer systems to make our cities more efficient. However, the Quantified City should focus on the needs of its citizens rather than the functionality of its technology when deciding what to measure, and therefore what to manage.
From Open Data to Actionable Knowledge
There is an important difference between data and knowledge. Data must be interpreted within specific contexts, and compared to relevant points of reference. Knowledge informs our decisions and behavior. Over the past few years there has been a strong push to increase the supply of open data, but our ability to transform such data into actionable knowledge has not caught up. It is for this reason that Socrata, one of the leading providers of open data portals for governments, recently introduced GovStat as a way to create performance dashboards from government datasets. The tool aims to give meaning to data by comparing performance over time and across geographies.
The GovStat platform took its inspiration from CitiStat, an award-winning governance program that began in Baltimore in 2000. When Martin O’Malley took over as mayor of Baltimore in December 1999 he wanted to know why the city suffered from such high rates of absenteeism. Within months he implemented CitiStat to monitor sick leave, overtime, and absenteeism in real time. During the first year of the program, the city saved $13.2 million and within three years overtime fell by 40% and absenteeism fell by up to 50% in some agencies. What gets measured gets managed.
Soon the platform expanded to track other performance indicators, from pothole abatement to trash collection, snow removal, illegal dumping, vacant buildings, and sewage overflows. Researchers that have studied the impact of CitiStat in Baltimore, stress the importance of strong support from the city’s leadership, beginning with Mayor O’Malley. Every week the head of each city agency was required to city down with O’Malley’s top staff to review the latest numbers from the CitiStat dashboard and adjust their practices accordingly. The data also helped inform new policies, such as a new waste removal program that increased recycling by 53%.
When O’Malley became governor of Maryland in 2007, he implemented StateStat a public-facing platform with clear, quantifiable goals and progress reports that admit when insufficient progress is being made. (Maryland is on track to reduce violent crimes by 20% by the end of 2018, but isn’t making enough progress to double public transit ridership by the end of 2020.) Notably, after O’Malley left Baltimore’s mayoral office, the city stopped regularly posting CitiStat data. Without committed leadership from the top, the data rarely translate into actionable knowledge.
Martin O’Malley deserves the recognition he has received for making public not just the performance indicators from government departments, but the progress toward specific goals that represent the interests of citizens. However, we must also be cautious about the downside of data-driven governance. The award-winning serial TV drama, The Wire, which depicts Baltimore during the time of O’Malley’s mayorship, frequently criticizes the culture of data-driven incentives for public workers. In the eyes of the shows’ writers, CitiStat caused teachers and police officers to “juke the stats,” but didn’t improve the quality of teaching or policing:
Indeed, in 2010 more than a 100 retired New York Police Department captains said they felt pressure to manipulate crime statistics in order to show reductions in reported crimes under the city’s CompStat program. (Though Heather MacDonald, writing in City Journal, questions the survey results and notes that crime re-classification dropped from 4.4% in 2000 to 1.5% in 2009.)
Regardless of the extent to which public workers feel compelled to alter data in order to achieve goals, we must recognize that quantified indicators are merely proxies for complex phenomena that our difficult to describe in numbers. A quality education is much more than standardized test results and the quality of healthcare at a hospital can’t be reflected merely in readmission and mortality rates. Still, these numbers — if collected and reported accurately — serve as important proxies that empower citizens to make better decisions and to ensure that public services meet the needs of those they are meant to serve.
The Quantified City is built on data, but ultimately comes down to communication. It uses numbers as a vehicle to promote conversations. Where are our cities headed? What are our needs? How should we allocate our resources? Imagine a city government that doesn’t just share its performance indicators and goals with the public, but invites them to monthly, thematic town halls to discuss what should be measured and what can be improved. Imagine a city government that doesn’t just track its own data and statistics, but also incorporates indicators from social platforms like Foursquare, Nike+, and Waze. Imagine a city environmental agency that works with high schools and universities to track water, noise and light pollution and then work together on reduction solutions. Not only is this all possible today, but it is affordable and technically unsophisticated. Much like the self-monitoring of the Quantified Self, all it requires is commitment and persistence.