Policy Recommendations for Implementing Civic Analytics

By Jessica Gover • July 30, 2018

From improving public health outcomes to supporting safer transportation systems, analytics can offer new insights and improvements to city governance across core issue areas. While the structure, application, and processes used to develop, iterate, and implement the Civic Analytics Network cities’ projects vary significantly, there are common policies adopted and operating within each that other cities can replicate. There is a need for more research and literature on city-level analytics to help practitioners and theorists alike understand how data-driven decision-making practices are operating within municipal governments. Our recent report, “Analytics in City Government: How the Civic Analytics Network Cities Are Using Data to Support Public Safety, Housing, Public Health, and Transportation,” seeks to contribute to that developing field and highlights steps that cities can adopt to develop their own analytics initiatives, catalogs ten examples of city analytics projects in the domains of public safety, housing, public health, and transportation, and, in closing, offers six policy recommendations to help other cities develop their analytics capabilities.

The “Analytics in City Government” policy recommendations, reproduced below, are derived from the Civic Analytics Networks’ work in public safety, housing, public health, and transportation and are offered to help spur and guide the effective development of analytics projects in other cities across the U.S.

1) Produce an Open Data Policy Roadmap

Adopting an open data policy can be a boon to rapidly, transparently, and collaboratively developing comprehensive analytics projects. Open data policies and portals enable city governments to operate with greater transparency to the public and to connect them directly to external researchers, algorithms, and/or datasets that can support more effective analytics project development. If advanced data use is a few steps down the road for a city, crafting an open data roadmap can help city government become more data-savvy to build towards future open data policy conversations. In early 2017, the Civic Analytics Network published an open letter to the open data community offering guidelines to help advance government capabilities for data portal development and to help deliver on the promise of transparent governance.[1] In June 2018, Civic Analytics Network members reaffirmed their commitment to those eight guidelines, publishing a one-year later letter to the open data community.

2) Create Programs and Job Descriptions that Promote Broad Data Literacy

By creating job descriptions and new frameworks for programs to appeal to data scientists considering positions outside of government, cities can attract dynamic, data-literate personnel to embed and distribute data skills at various levels within city government. Whether it is creating a CDO position, establishing an analytics team, or simply embedding a data scientist within a department, establishing a role for data expertise with the support of senior leadership can enable city officials to develop the analytics projects that they need most. In addition to bringing in data champions and expertise, cities can also improve internal capacities by designing training programs to provide critical professional development opportunities to city hall personnel. San Francisco’s SF Data Academy which provides a pathway for city employees to receive continuous professional development focused on data skills and analytics from within government, is a leading example in city-level analytics training programs.[2]

3) Incentivize and Enable Cross-Departmental Collaboration to Connect Personnel and Data Resources from across City Government

Developing an analytics project places data scientists in an internal consultant role, as they are typically situated outside of the department where that project will be implemented. It is important to establish pathways for data scientists to collaborate and receive input from the relevant department or agency, and, simultaneously, pathways to incentivize city personnel to engage with those new data science experts need to be established by supervisors or even from the bully pulpit. Government personnel operate in a bureaucracy and it is important for their supervisors to establish space in their day-to-day responsibilities to help them ‘make the time’ to engage on data-focused projects. By incentivizing department or agency personnel to connect with data scientists within city hall, supervisors or even the bully pulpit can open up space within city staff’s day-to-day schedule and performance requirements to enable them to pursue analytics solutions to core issues. Beyond connecting departmental personnel to data scientists via substantive pathways for collaboration, cross-departmental engagement is also key. While many city departments maintain useful data repositories, that data is often siloed or incompatibly structured, rendering analyses with data maintained by various departments infeasible. Establishing resources, tools, or policies to help streamline data standardization and warehousing can enable cross-departmental data sharing and is a critical facet of becoming a data-smart city. In Los Angeles, GeoHub, the city’s open data platform, offers unprecedented access to the city’s highly integrated data resources. GeoHub is a publicly available platform designed to allow the public to explore, visualize, and download location-based open data. It also allows departments across the city to share, access, and collaboratively utilize other departments’ data.[3] Other Civic Analytics Network members are working to replicate this platform in their cities.

4) Adopt Enterprise-Wide Procedures that Facilitate Data-Driven Insights

Whatever methods a city uses to encourage data analytics, adopting an effective project management process means establishing a policy framework that enables data science experts to design analytics projects with the support of the city’s legal, administrative, and oversight capacities. “Human-centered design” is a commonly used method among Civic Analytics Network cities and offers methods that help craft analytics projects that are responsive, equitable, transparent, and designed with community members in mind. To develop useful analytics projects, cities need enterprise-wide procedures, such as data usage practices, security protocols, or standardized legal and data sharing agreements. For example, in New York City, the Mayor’s Office of Data Analytics (MODA) created the MODA Process Map to help departments develop data use practices and internal awareness.[4] While these procedures can help streamline, stabilize, and embed data use practices across government, project managers must be mindful of potential blind spots, such as algorithmic biases, that may be unwittingly built into their models. Allegheny County, Pennsylvania, a county-level member of the Civic Analytics Network and home to the City of Pittsburgh, a network member, has developed a Data Warehouse to create a more efficient and data-driven environment for the delivery of human services. The Data Warehouse has enabled county administrators to learn more about individual clients and address gaps in coverage.[5] Many Civic Analytics Network cities are now pursuing their first data warehouses, and this network-wide trend represents an important shift in practice from project-based data-use efforts to organization-wide strategic data practices and policies.

5) Link Civic Engagement with City Analytics

Chief data officers may work within the walls of city hall, but they are members of a broader community and data ecosystem. The best analytics insights come when city government data use and civic engagement converge—after all, the public is the constituency for city analytics. Whether an organization is analyzing datasets available on an open data portal, developing a data visualization, or scoping a predictive analytics project, the results any of these efforts yield are better crafted when co-created with the public. By producing analytics models informed by direct input from city residents or developed in partnership with a civic tech group, cities will garner better service improvements and data-driven insights.[6] Kansas City, MO, for example, uses a quarterly feedback mechanism called the Citizen Survey through which residents can both respond to prompt questions provided by the city and communicate their top priorities for the city.[7] Citizen Survey is a leading example of a municipal citizen survey tool and has established a continuous feedback loop to link citizen perspectives to Kansas City’s performance and services.

6) Produce Guardrails to Protect Equity and Fairness Issues

Analytics is a practical tool for overcoming resource shortages and for distilling vast and disparate data, but it can also lead to the reproduction of biases and inequities under the banner of data science. Establishing standards of practice and mechanisms that ensure clear and continuous engagement with the public are critical components for cities to maintain transparent, equitable governance, and for incorporating inclusive analytics practices into city government. Under the leadership of San Francisco’s chief data officer and the new chair of the Civic Analytics Network, Joy Bonaguro, the network is developing a toolkit to help cities assess the risks and biases of algorithms. This toolkit focuses on algorithms developed both within city hall and by third-party vendors, and aims to help safeguard city analytics so that future data-driven efforts are able to produce fair and equitable solutions for the benefit of all community members.[8]

The full report, "Analytics in City Government," can be found here.

[1] See “An Open Letter to the Open Data Community” (Ash Center) https://datasmart.ash.harvard.edu/news/article/an-open-letter-to-the-ope....

[2] See “San Francisco’s Data Academy Develops a Data-Savvy Workforce” (Ash Center) https://datasmart.ash.harvard.edu/news/article/san-franciscos-data-acade....

[3] See “The Power of Data Visualization in Cities: Los Angeles’ GeoHub” (Ash Center) https://datasmart.ash.harvard.edu/news/article/webinar-the-power-of-data....

[4] See “Allegheny County, Pennsylvania: Department of Human Services? Data Warehouse” (Ash Center) https://datasmart.ash.harvard.edu/news/article/allegheny-county-pennsylv....

[5] See “Mayor’s Office of Data Analytics (MODA) Project Process” (New York City) http://www1.nyc.gov/assets/analytics/downloads/pdf/MODA-project-process.pdf.

[6] See “How Citizens See It: Kansas City’s Citizen Survey Adds Citizens’ Perceptions to the Equation” (Ash Center) https://datasmart.ash.harvard.edu/news/article/how-citizens-see-it-677.

[7] See “Customer-Driven Government” (Ash Center) https://datasmart.ash.harvard.edu/news/article/customer-driven-governmen....

[8] Also See “Potholes, Rats, and Criminals: A Framework for AI Ethical Risk” (Ash Center) https://datasmart.ash.harvard.edu/news/article/potholes-rats-and-criminals.

About the Author

Jessica Gover

Jessica A. Gover is a Ph.D. candidate in Political Science at Johns Hopkins University and a writer for Data-Smart City Solutions. Previously she conducted research on new approaches to tech-enabled innovation in the U.S. federal government, focusing particularly on the executive branch, as well as innovative public-private partnership models in the U.S. and New Zealand. Her work for Data-Smart City Solutions is informed by research interests ranging from democratic governance to regulatory policy to civic engagement to cross-sector partnerships. She holds a Master's degree from the University of Chicago and received her B.A. at Trinity College.