Information: The Key to Effective Government

• July 28, 2014
• Civic Data

This is the second in an article series based on Stephen Goldsmith's paper "Digital Transformations: Wiring the Responsive City." Click here to read the report in full.

Government can perform most tasks reasonably well—picking up garbage, responding to emergencies, funding programs—without high technology. In New York City, for example, the police perform superbly, sanitation workers rarely miss a pickup, and firefighters are on hand for every emergency, always well staffed and brilliantly prepared. Workers deliver these good results despite layers of supervision, lawsuits, suspicions, media aversion, labor contracts, job classifications, and other constraints (constraints that help explain why the city’s workforce—350,000 people, counting teachers—is so massive). But vast armies of able government workers can do only so much. If not managed well, these armies duplicate work, interfere with one another, and create extra rules and paperwork that sap government energy and focus. The key to effective management is the efficient treatment of information.

All organizations depend on three basic processes. They need to get information (data), they need to make sense of that information (analysis), and they need to act on that information (action). The nature of reform becomes clear once one sees how responsive government handles each of these three essential aspects of information management—all quite differently, in fact, from the traditional bureaucratic approach.

The Battle of Simplicity and Flexibility

In the past generation, two paradigms of information handling have battled for the soul of government. One paradigm is the structure of bureaucracy invented in the late nineteenth century and enshrined in that era’s ideal of progressive government. Developed to bring cities out of the stagnation of corruption and nepotism, this bureaucracy runs on a hierarchical structure with a detailed division of labor, guided by formal rules and regulations. Its aim is to make each job simple to perform and to evaluate.

Progressive reformers of the late nineteenth century saw that accountability required a reduction in discretion, and they certainly succeeded in that goal. The method they used, though, was the elimination of discretion altogether. In the standard bureaucracy of the last century, oversight systems imposed by law—civil service and union contract work rules, job classifications compounded by inspector-general investigations—reinforced the importance of the narrowly defined routine. In this first paradigm, cities “manufacture” government with more brute force than finesse. And when they need more government, they simply purchase more units of input: more labor, more inventories, and more trucks, for example.

When a police officer’s job is to respond to a dispatcher, a teacher’s to conduct a class, and a public nurse’s to perform checkups, classic bureaucracy—with its clear assignment of responsibility to individuals to perform discrete tasks—works well. But we now live in a digital world where work does not stay within such neat and discrete divisions. It is now quite possible for employees in separate locations and agencies to work on the same file concurrently, for supervisors to see “red flags” and outliers in real time, and for wireless devices and data analytics to create new forms of supervision. And when an employee’s work requires sharing information and making decisions not specified in job descriptions, the simplicity of old-style bureaucracy becomes a liability. It is then that we see the importance of the second governmental paradigm: the need for flexibility in the face of complex problems.

In traditional bureaucracy, a dilapidated building is a matter for people whose job it is to deal with construction and maintenance: the inspectors who ensure that work is done in compliance with city codes. In today’s city governance, the building is correctly seen as a concern for workers in public health, police, environmental affairs, economic development, and other areas beyond building-code enforcement. It is simply more efficient and cost-effective to grapple with all the different causes of a building’s broken-down state. Understanding those factors could yield the best strategy for repairing the building. Such a complete analysis can reveal that fixing a problem requires a multipronged approach; other times, it shows how focusing on one key element will tip all the others.

To respond with this kind of flexibility, a post-progressive bureaucracy must allow professionals the discretion to explore the complexity of policy challenges and, working together, to design collaborative responses to those problems. This is the ideal of today’s reformers.

Innovation Infrastructure

Ironically, the data-driven revolution of today’s reformers has its origin in pushpins on paper maps. An officer in New York City’s Transit Police Crime Analysis Unit used the pins to designate where crimes occurred. The unit was then able to visualize clusters and patterns, enabling them to precisely focus their efforts on trouble spots. In the mid-1990s, the pushpins gave way to CompStat, a more sophisticated data collection, analysis, and mapping system adopted to manage the New York Police Department. CompStat, which quickly fostered a 60 percent reduction in crime, inspired Baltimore’s CitiStat, a performance-management system that extended the concept of data awareness and flexibility beyond policing.

Through the first seven years of the program, CitiStat helped Baltimore save about $350 million by improving the efficiency of its departments. The system was later adapted to work on a statewide level after Baltimore’s mayor, Martin O’Malley, became governor of Maryland in 2006. In today’s Maryland, data and analytics combine to make government responsive in ways that scarcely could have been imagined 20 years ago. For example, algorithms predict the risk that a certain parolee or probationer will reoffend. That person will then receive increased supervision.

Such techniques allow the state to utilize scarce resources where they will have the most impact. This management approach is now spreading to other municipalities, where it is usually associated with improved services and lower costs. For instance, in Louisville, Kentucky, under Mayor Greg Fischer, the city’s asthma-monitoring project uses GPS data from inhalers to identify and mitigate the environmental conditions that cause chronic asthma.

For all their differences, these and other responsive government efforts share several key traits. They all involve returning discretion to municipal employees, informing that discretion with vast amounts of data, and analyzing data in ways that help civil servants solve problems. At the same time, these new approaches give citizens a clear window into the operations of their government. Importantly, too, all these efforts have involved new management structures and have had the support of key executives determined to change “business as usual.” A look at three major cities shows how these themes play out in different ways in different municipalities.

Chicago: How Centralized Data Analytics Change the Essence of Government

When people mingle socially in a single place, they organize and synthesize information. Hence data analytics centers, where previously separated streams of data are collected in one place, are now among the most powerful drivers of innovation in government. Such centers do much to solve the biggest problem in government today: isolation of knowledge and resources from problems. These efforts drive change by mining data that power predictive analytics, which can then be delivered to the field through decision-support tools.

A leading example of the power of data centers is the work of Chicago’s Department of Innovation and Technology (DoIT), which is transforming that government’s data from an after-the-fact archive to a resource that guides proactive decisions. DoIT has the sort of control over data collection and analysis that used to belong to individual departments. The DoIT approach lets ordinary citizens access information that they could never see before: Chicago now has one of the largest and most robust open-data portals in the United States, with 592 data sets, ranging from the salary of every municipal employee, to every reported incident of crime from 2001 to the present. Chicago’s SmartData platform soon will be able to pull information from across the government, facilitate predictive analytics, and display results on an interface that lets users see data geographically in real time. Another DoIT component of SmartData is the geographical interface—called WindyGrid—which allows users to view geographic and temporal-based data mapped across the city.

WindyGrid was developed in early 2012 by then–chief data officer (CDO) Brett Goldstein, who envisioned a program that would be able to tell the story of a given location at a given time. Thus, by querying a specific location in the city—such as the intersection of State Street and Madison Street—a user would be able to see a history of 311 calls, 911 calls, crime reports, and other occurrences at State and Madison, as well as any activity occurring there in real time.

SmartData’s predictive analytics capability will enable city officials to observe relations among disparate sets of data, providing estimates of the probability of one type of event occurring after another. For example, the system revealed that concentrated rat populations tended to spike seven days after garbage-related complaints—thus allowing the city to mobilize forces accordingly when garbage-related complaints occur. These analytics would then be displayed on the WindyGrid interface, providing users with a powerful tool for coordinating incident management and workflow.

Chicago’s innovations have succeeded in no small part because they had the strong support of Mayor Rahm Emanuel and the concentrated resources required to succeed. For instance, Chicago developed its own, in-house “center of excellence” around data analytics, hiring the first-ever CDO in Brett Goldstein, whose mission as CDO was to coordinate the city’s open-data policy and lead new data analytics initiatives. Throughout his tenure, Goldstein oversaw the expansion of Chicago’s data portal and used innovative new methods to use the city’s data reserves. For example, he developed WindyGrid using open-source tools at a small fraction of the cost of traditional software procurement.

Under the leadership of CIO Brenna Berman, DoIT has been developing its internal capacity for innovation and product development. The department currently administers the growing WindyGrid application, is in the process of developing the SmartData Platform, and has plans to design other resource tools as well. By making Chicago’s data transparent, DoIT has been able to leverage the talents of the city’s growing community of “civic hackers,” techies who use their programming skills to develop meaningful apps in the public interest.

Furthermore, Chicago’s open-data policy enables academic participation in problem solving. DoIT has formed essential partnerships with the University of Chicago and with Carnegie Mellon University, both of whose considerable capabilities are a component in Chicago’s development as a center for data analytics.

Thus, through strong executive leadership, the cultivation of talent for in-house development, and key strategic partnerships, Chicago and DoIT have been able to set the stage for creating a new, data-driven policy framework that was unthinkable only several years ago.

Boston: Structures That Change the Citizen’s Relationship to Government

In the standard bureaucratic model of twentieth-century government, citizen activists often assume that they need to overwhelm local officials with information about how common problems are. When I served as New York’s deputy mayor, one community board leader told me: “I always assumed that the volume of complaints determined whether city officials would respond.”

But the availability of easily accessible 311 data, she went on, had increased her effectiveness as an advocate. Instead of looking for ways to make noise, she now looks for patterns in information. Once she finds such a pattern, she can ask the government to craft an appropriate solution. The availability of online data helps people figure out the causes of problems rather than assuming the answers without substantiation. As an example, the board leader told me about a series of accidents involving elderly pedestrians at a traffic intersection. Publicly available data about those mishaps led her to realize that they arose from insufficient time to cross the street to a pharmacy—meaning that timing changes in walk/wait signs could address the problem.

This new model of engagement cuts across agencies. Safety problems might result from policing inadequacies; but they might also result from the design of streets or parks, activities of nearby schools, schedules of delivery at nearby stores, or any other factor. So data liberate public officials to do what they do best: identify problems and develop solutions—not just in their own department, but by working with people across government.

Boston’s responsive government effort stresses this kind of engagement with the public. The city led the country on this front in 2010, when it created a specific mayoral office dedicated to service delivery and citizen engagement: the Mayor’s Office of New Urban Mechanics (MONUM). The office lies within the mayor’s suite (another example of strong support from the top), and its cochairs come from backgrounds mixing private-sector experience with public-sector interests. The MONUM incubates innovative projects and facilitates collaboration within city hall and with outside entities. “Civic innovation is civic engagement,” says Mitch Weiss, the former mayor’s chief of staff. The unit’s name is a nod to the fact that former mayor Thomas Menino has long been called—both admiringly and critically—an “urban mechanic.” Yet no one knew the city and its neighborhoods better. The MONUM is thus an amalgam of new technology and old-fashioned attention to the details of urban life.

The MONUM illustrates how responsive government needs a “safe space” within city hall for experimentation, and it allows the city to be ambidextrous: to have city agencies focus on the traditional work of service delivery while allowing “citizen mechanics” to experiment on new initiatives. For example, Citizens Connect, a smartphone app, allows citizens to report problems throughout the city. The reports, usually accompanied by a photo taken with a smartphone, are tied to specific locations in the city. Citizens report issues such as illegally parked cars stopping traffic, rat infestations, building elements or trees that may fall onto or impede pedestrians on the sidewalk, graffiti, or trash left out; all these reports are then automatically mapped and sent to the government to sort out. Such complaints deal with issues touching the full range of departments—police, sanitation, traffic, housing, neighborhood services, and parks. When collected and organized in the appropriate databases, these reports can serve two purposes. First, they can serve as work orders for city workers. Second, they can help identify patterns: “hot-spots” where broader initiatives and policies need to be initiated. This has been so successful that the state funded a rollout of Commonwealth Connect, which has brought this functionality to more than 40 communities in Massachusetts.

Digital platforms allow government to quickly capture community sentiment and translate that information into action. Citizens generate data when they converse with their governments via social media, when they participate in online “ideation” forums, or when they use city apps to report problems or rate services that they receive from government departments. Government can curate and use these new streams of information to operate more effectively and bring in innovative ideas from the community.

Boston’s success with citizen engagement reflects a broader trend evident in many other municipalities. In 2012, for instance, Philadelphia formed its own Mayor’s Office of New Urban Mechanics through an executive order placed by Mayor Michael Nutter. Meanwhile, Mayor Vincent Gray of Washington, D.C., provides an innovation infrastructure in a more direct retail way by providing a platform for collecting and visualizing customer grades of city services and employees. Grade.DC.Gov solicits, mines, measures, and displays resident satisfaction and reviews, placing emphasis on the quality of government services as perceived by the citizens being served. Consumers of city services can use the Grade.DC.Gov website itself or provide input by texting, posting on blogs, or using social media services (such as Facebook, Yelp, and Twitter). A daily digest helps city officials start their workdays with a list of comments from the previous 24 hours, allowing for quick remediation of problems. Use of Grade.DC.Gov has resulted in departments improving their grades consistently from the outset—from “barely passing” to As and Bs. These innovations illustrate the power of new technology to improve responsiveness and communication between citizens and their government.

New York: Innovation Delivery Units Meet the Digital Scientist

All the themes of responsive government—data freed from separate silos and made available to all, powerful analytics, executive commitment, and the active quest for useful innovation—can be found in New York City’s recent efforts. You would expect a government led by former mayor Michael Bloomberg to care about data. Before going into politics, after all, Bloomberg made his fortune by selling terminals that displayed comprehensive, minute-by-minute data to stock-market traders. In fact, Bloomberg’s New York measured just about everything, from 311 call trends to tree plantings to detailed, agency-specific indicators. Mayor Bloomberg’s famed “bullpen” in City Hall—designed to emulate the trading pits on Wall Street, where information and energy flow freely—contained a large screen that displayed performance data all day. The information flow encouraged officials to think not just about specific problems, but also about the connections among city departments.

Hundreds of indicators helped the mayor and his team focus on areas needing attention. Of course, streams of real-time data have their limitations: they tend to measure activity more than outcomes; they are retrospective looks at the past rather than predictions; and they often focus on a single agency. But such data streams are a critical first step toward a new and broader digital effort—one that combines and analyzes information across agencies to find and preempt problems. In New York, an ad hoc data-mining group (a self-described “skunk works” in City Hall) stepped up to fill the gap between raw data and useful knowledge, using analytics to identify answers to important questions.

When I served in New York City government (2010–11), we discussed three models of how to drive digital innovation. One model was to establish a digital analytics center within the mayor’s office, with links to each agency. Such a center would look for discoveries that would transform the way government works. As we conceived it, the new revenues discovered from underreported taxes, waste, fraud, and abuse would fund important new work in social services and other areas. At that time, though, our budget office could not precisely define the benchmarks that would have let us define substandard performance. The city did not create the formal analytics center. Instead, the government used a second, more ad hoc model, empowering a smart group of data scientists who used information to unlock discoveries across agencies.

Our third model came from social-service agencies that had concurrently started their own data exchange, HHS-Connect. Rather than seeking to connect siloed data sets by assembling them into a large data warehouse, this initiative took another approach. It linked disparate information sources via information portals, where once-separated processes were now combined. For example, because of HHS-Connect, a citizen applying for public benefits need not resubmit a birth certificate that had already been furnished to the city for another purpose. This makes dealing with the city easier for citizens, while helping social-services employees improve the quality of casework by giving them access to additional relevant data about customers.

An early example of New York’s success in responsive government was motivated by a fire on April 25, 2011, in a row house in the Bronx. The blaze killed a 12-year-old boy and his parents. More than 100 firefighters worked for an hour to contain the fire. The boy’s family was one of several who lived in illegal conversions in the row house. As early as 2008, neighbors had called the Police Department, the Department of Buildings, and the 311 phone line, complaining that the building housed illegal single-room occupancy units and did not, among other violations, have appropriate exits. So why hadn’t anyone responded?

The building and fire commissioners, in fact, both had information about the building. Yet they also had an overload of data about structures all over the city, and the inspectors whom they sent to investigate complaints were rebuffed by the building’s tenants. None of their data sets were organized in a way that helped either department identify priorities or ways of pursuing those priorities with other city agencies. The firefighters were doing their best just to keep up—they were more likely to respond to crises than to complex, simmering problems.

But what if the fire department’s goal was not just to respond to crises but to predict what problems could lead to crises? What if fire officials could identify the causes of the fires—before the blazes struck? What if the city could predict where a fire is most likely to occur? This information would allow the city to strategically target its resources to inspect and evacuate the most dangerous dwellings. To pursue this goal, the city launched a pilot project guided by Mike Flowers, a lawyer and data analyst, which focused on fire risk and used metrics to identify the most dangerous of the illegally subdivided properties. The Flowers-led effort used the following metrics to identify the properties most at risk for fire:

• Owners in financial distress (including those with foreclosures or tax liens)
• Multiple illegal-conversion complaints
• Multiple-family dwellings built before 1938, when a significant building-code revision took effect
• Low-income/high-immigrant/low-employment neighborhoods

The special team consulted easily available data, such as agency reports, real-estate filings, and finance and tax information. The analysts found that dwellings with all four risk factors were over 40 times more likely to have a fire. Further screening allowed the team to target true high-risk illegal conversions, attaining a 40 percent vacate-order rate for single-room-occupancy conversion complaints, up from a historical three percent. The project showed that, by intelligently moving from a focus on activity to a focus on cross-agency data to achieve specific outcomes, resources can be much more effectively deployed. And, as with most comprehensive review efforts, the study’s insights led to other discoveries. For example, what about the illegally subdivided buildings that had a low risk of fire? In a place like New York, which has a severe housing shortage, perhaps zoning or other regulatory changes should be made to allow such conversions. We can see from the New York experience that a traditional data-tracking and management tool, which simply evaluates the elapsed time to an inspection, would not produce the same result as identifying and mitigating risk. The next step, in the evolution of new structures to power data innovation, takes the New York model into an institutional solution for a city.

Conclusion

The examples of responsive government described above have several important, common traits. First, they are all real: practical applications of innovative thinking and new technology, driven by civil servants committed to change. Second, they are all successful: these programs have found solutions to stubborn problems, replaced the heat of conflict with the light of information, and saved millions of dollars by freeing information from its former constraints. Third, they use technologies that consumers increasingly expect from private industry, such as Netflix, Amazon, Apple, and Google.

These traits illustrate some of the means and motivation for responsive government. An easy exchange of ideas and technology with the private sector helps provide government with the technological means. And now that consumers see how well the private sector uses data intelligently to meet their needs, they increasingly expect government to do the same. That is the political motivation: urban residents accustomed to intelligent use of data in the private sector will not long tolerate “data blindness” in city services. The era of responsive government is here to stay.