Jane Wiseman Grey

By Jane Wiseman • January 25, 2017

This paper is part of the Civic Analytics Network initiative at Harvard's Ash Center. To download this paper as a PDF, please click here.

Executive Summary

A Chief Data Officer (CDO) can lead a city or state toward greater data-driven government. Leveraging data enables more responsive and rational allocation of government resources to address priority public needs. Data-driven executive leadership in government is relatively new, with just over a dozen cities and a handful of states having named a CDO as of late 2016. There is growing momentum and increasingly frequent news of the next government CDO appointment. While there is a growing proliferation of CDOs in government, there are few resources that describe the landscape, either for the benefit of the chief executive appointing a CDO or the new CDO taking office. This paper intends to help new entrants by documenting selected current practices, including advice shared by existing government CDOs, observations by the author, and analysis from government technology and analytics experts. A few key points for a new CDO to consider include:

  • Support from the chief executive sets the CDO up for success. Whether a CDO reports to the chief executive (mayor, governor, or county commissioner) or not, it is important to have the support of that chief executive and have the resources, credibility, and authority that go along with executive sponsorship.
  • Basic management skills can accelerate progress. Strong basic management and leadership skills, the ability to clearly articulate the mission and roadmap to achieving it, and the ability to hold staff accountable for results will accelerate success for a CDO. Standardizing tools and processes, including project management tools, will make the work more efficient. Balancing the demand for results with the need for foundational data stewardship demands leadership from a CDO and a delicate balance of people and technology skills.
  • Data stewardship can create the conditions for solid analytics. Data stewardship – comprising data governance and data infrastructure – lays the foundation on which analytics is built, and whether these activities are part of the CDO operation or not, they are essential to the success of any analytics program.
  • Setting priorities becomes an increasingly important and challenging task for a successful CDO. As the profile of the CDO grows and demand for services increases, it can be difficult to manage priorities and stay true to the mission. As one expert advised, CDOs should stay focused on important policy issues and operational improvements in government, and avoid “data qua data” analytics.

This paper describes an operational framework for the role of a CDO office, and provides observations on fostering a data culture in government.


The Civic Analytics Network project, including this paper, is generously funded by the Laura and John Arnold Foundation. This paper reflects the input of many individuals. Valuable contributions came from the individuals listed as interviewees in the sources section of this document, without whom the paper would not be possible. The advice and guidance of Stephen Goldsmith cannot be underestimated. The paper would not have been possible without the keen editorial advice and research support of Craig Campbell and the substantive feedback and editorial support of Katherine Hillenbrand, as well as insightful feedback from Zachary Markin. Data fellows who contributed to written material about their cities include Robert Burack, Sari Ladin, Sean Thornton, and Blake Valenta. Note that some comments and advice were contributed by individuals who wished to remain anonymous. This report is an independent work product and views expressed are those of the author and do not necessarily represent those of the funder.


Within the past year, a number of cities and states have created the position of Chief Data Officer (CDO) to lead their efforts toward data-driven government. There is wide variation in the responsibilities of CDOs – some focus on open data programs, others focus on applying geographic information systems (GIS) tools to government problems, some preside over performance management, and still others are taking on advanced data analytics projects. The majority of new CDOs are appointed as the first and only employees on their team. A few start with a team already in place, and others are able to hire soon into their tenure.

Regardless of the size of their team or the scope of their responsibilities, CDOs share a common goal – to increase capacity for making data-driven decisions in government. Ideally, the work of a CDO results in greater value for the taxpayer dollar through better allocation of existing resources to meet the needs of the public.

In 2011, the city of Chicago appointed the first municipal Chief Data Officer (CDO). That same year, New York City Mayor Michael Bloomberg charged a team with a specific analytics task, and by 2013, that group became the Mayor’s Office of Data Analytics and began to take on applying analytics to citywide operations. These two pioneering cities along with others including Boston, Los Angeles, New Orleans, Philadelphia, Pittsburgh, and San Francisco created the earliest CDO offices and analytics programs from scratch using creativity and hard work – and no preexisting roadmap for a municipal CDO. The same is true for state CDO offices such as in Colorado, Connecticut, and New Jersey.

The CDOs in early-adopter cities created an informal network for peer support and sharing of best practices in 2014. In 2016 the group formalized as the Civic Analytics Network, supported by Harvard Kennedy School. The effort includes collaboration on shared analytics projects among network members with the Center for Data Science and Public Policy at the University of Chicago. Stephen Goldsmith, Daniel Paul Professor of the Practice of Government at Harvard Kennedy School, facilitates network convenings and provides thought leadership.

This paper documents operational insights generously shared by Civic Analytics Network participants along with observations from other CDO offices and external research sources. Describing the experience of pioneering CDOs is intended to make it easier for those who follow in their footsteps. Just as the steam engine powered the industrial revolution, the creation of offices of data and analytics can power a new revolution in how government delivers services to the public by turning data into actionable insight.

Appointing a Chief Data Officer

State, county, and city executives have named a new Chief Data Officer (CDO) nearly every month in the second half of 2016. The observations of experienced CDOs in this paper can offer insights that can create the conditions for success for a new CDO.

Senior Executive Sponsorship Sets the CDO up for Success

To be successful, a CDO must have a clear mandate from and the support of the chief executive. While a minority of city CDOs report directly to the mayor, having his or her support is important for success. One way a chief executive in government can demonstrate commitment is by issuing an executive order or policy statement about open data or the creation of the office of the CDO. Some mayors and governors have been quite public in their expressions of interest in data-driven government.

For example, Los Angeles CDO Lilian Coral credits the mayor’s keen attention for her success in launching the open data performance dashboards as well as the easy-to-use public mapping site GeoHub, which has increased both internal and external access and understanding of the city's open data. New Orleans analytics leader Oliver Wise credits his mayor’s commitment to data-driven government with giving his office the authority to work as a trusted partner with city departments, enabling accomplishments ranging from reducing blight to increasing fire safety. Boston’s CityScore, a single citywide numerical score to measure the performance of key city services, has been a high priority for the mayor, and his sponsorship was critical to securing cooperation across departments for the data inputs.

A passionate sponsor who can clearly articulate the value of open data, data visualization, and analytics and can make that value concrete for both municipal employees and the public is an essential advocate for data-driven government efforts.

Who Does the CDO Report To?

Often, particularly in larger cities, Chief Data Officers in government today report to a Chief Information Officer, as is the case in Boston, Chicago, Minneapolis, Pittsburgh, and Washington, D.C. Some are embedded in the same organization as the performance office, as is the case in New Orleans, Louisville, San Diego, and Seattle. Some report to their mayor or to a deputy mayor, as is true in San Francisco and Los Angeles.

CDOs reporting directly to a mayor may be more likely to focus on projects with high visibility to the mayor and to the public, and as with other direct reports to a mayor, may operate under pressure to produce visible results quickly. Those CDOs not directly reporting to a mayor are more likely to be able to take on projects requiring more time to demonstrate results, including more methodologically complex analytics projects.

Comparing the CIO and the CDO Roles

Where there is already a Chief Information Officer (CIO), as is true in most big cities and in all states, naming a CDO provides a blank slate of opportunity – a chance to define a role with strategic priority to the chief executive. The role of a CIO can remain operational, keeping large IT projects moving forward and supporting enterprise-wide systems and platforms.

This opens the way for a CDO to take on the emerging and dynamic strategic role of defining how data can be used to create intelligence to improve performance. The CDO can unlock the value in data that the CIO stewards into existence.

Operations vs. Strategy: Key Differences Between CIO and CDO Roles


Key Data Responsibilities

Key External Partner Responsibilities


Build and maintain enterprise data systems that produce large volumes of data

  • Run procurement process and select vendors
  • Manage vendor relationships
  • Identify and develop external pro bono partnerships


Use data in describing, mapping, and modeling to identify patterns and to derive business value and policy insight

  • Work with contracted vendors
  • Identify, develop, and sustain external partnerships by working collaboratively on analytics projects. Partners can include the civic tech community, business partners for pro bono work, and academic partners
What’s in a Name?

While Chief Data Officer is the predominant title given to leaders responsible for managing data and analytics in government, it is by no means the only title. Other titles given to this role include: Chief Analytics Officer, Analytics and Strategy Manager, Data Analytics Lead, and Director of Enterprise Level Data Management. As the most common title, CDO is used in this paper to refer to the portfolio of activities that create a data culture and enable greater use of data to drive decision-making.

Even the teams led by CDOs have a variety of names – in Boston it is the Citywide Analytics Team, in New Orleans the team is called NOLAlytics, in New York City it is the Mayor’s Office of Data Analytics, and in San Jose it is the Data Analytics Team.

Does a CDO Have to Be a “Chief”?

A CDO will be more effective if he or she has the authority to lead, and to bring departments together for shared projects. Being of sufficient seniority to have peer-to-peer discussions with key operational leaders across government helps a CDO lead cross-departmental and cross-disciplinary projects. The “chief” title also signals the importance of the role to the government’s chief executive.

Strategic Framework for a Chief Data Officer

The scope of responsibilities for a CDO varies and can include any combination of the following: open data, data collection and data sharing standards, data management, performance management, geographic information system (GIS) and data visualization, and data analytics. Regardless of the scope, the CDO is a leader for data-driven government and some aspects of the job are common across all CDOs.

Many government CDOs, particularly newer ones, are in their early stages and have not yet formalized their strategies and operating models. The graphic below presents a framework for thinking about the components of a CDO role, from developing a strategy, to building a team and engaging with departments, to conducting analytics and fostering a citywide data culture.

While presenting these steps in a linear graphic may imply that there is a strictly linear relationship among the responsibilities, they are more often iterative. For example, while strategy necessarily comes first, it should be revisited periodically. And while building the team happens early, ideally it is ongoing either due to the success and growth of the team or due to naturally-occurring attrition and turnover.

Some parts of the framework may seem like “management 101,” but they merit discussion in the context of the challenge of being a new CDO, as a desire for fast or demonstrable results could crowd out attention to the “nuts and bolts” of organizational operations. Each element of the strategic framework is described in the pages that follow.



The strategy phase includes setting a clear charter, developing a mission statement, and creating a roadmap for implementation, as described below.

Clear Charter

Strategy begins with the charge from the chief executive, which should answer key questions, such as: what role has the CDO been asked to play in government? Does the CDO report to the chief executive or is the CDO embedded in another part of government (IT, performance, innovation, budget office)? Is the CDO expected to take on the full scope of related tasks – open data, geographic information systems (GIS), and analytics, or will some subset of those tasks be in the charter? How does the CDO relate to other parts of government and how are responsibilities divided? What is expected on an annual and on a routine basis of the CDO by the chief executive (mayor, governor, county commissioner)? What resources have been provided to support the charter? Is the charter made clear across government in an executive order or other public document? How is the CDO being announced and introduced to peers within government?


With a clear charter from the chief executive, a CDO can define a mission, a concise statement of the difference the CDO can make in government over a three- to five-year time horizon. A mission statement is valuable in making clear to departments the types of support a CDO can provide, and just as importantly, the tasks that are out of scope.

For CDOs with a team larger than one, having a clear statement of mission provides direction to the team. Mission statements help a team prioritize, because everything should connect back to the mission. In a truly mission-driven organization, every team member knows the mission and can communicate it consistently to all stakeholders.

For example, the City of Chicago’s mission statement for the CDO office is “to use data to improve the quality of life of residents in the city and improve the efficiency of city operations.”[1] The New York City Mayor’s Office of Data Analytics’ mission is to be the city’s “civic intelligence center, allowing the City to aggregate and analyze data from across City agencies, to more effectively address crime, public safety, and quality of life issues. The office uses analytics tools to prioritize risk more strategically, deliver services more efficiently, enforce laws more effectively and increase transparency.”[2]

Implementation Roadmap

Building a roadmap lays out the steps and actions and the timeframe required to achieve them. A roadmap can take the form of a strategic plan or an implementation plan. A roadmap should identify the resources necessary to execute, including the skills needed and the process for identifying them. It should indicate the timeframe for each activity as well as who is responsible (staff, vendor, pro bono partner, etc.). It is helpful to update or revisit the roadmap on an annual basis.


Making a strategy or roadmap public is a great way to support both messaging efforts and also to build transparency. San Francisco has an excellent one-page summary of its data strategy. The summary contains the mission, vision, goals, and each element of the strategy. For each area of the strategic plan — services, tools, data repositories, outreach, and data governance — color coding tells the stage of completeness for each activity. At one glance, city departments can see what the San Francisco CDO and team are doing and what they plan to do. This helps city departments understand how the CDO can help them improve their operations.

Having the charter in writing, whether in the form of a mission statement, strategy or implementation plan, builds clarity for the CDO team as well as for the rest of government.  


Building a Team

A CDO is typically the first hire. Depending on the charge and the available resources, the CDO may be a one-person shop briefly or for an extended period. For CDOs who have the resources to build a team, the following pages address the issues to consider in creating the team.

Defining Roles

The size of a CDO team can be one, or a handful, or up to 30 in the largest operations. The roles and skills to be acquired in building the team will depend on the mandate and the scope for the CDO. Due to the wide variety of tasks asked of a CDO organization, generalists with strong analytical capabilities are often critical staff members. This is especially pronounced within the smallest organizations.

For a small team, each team member may wear multiple hats and take on a variety of tasks. Even in CDO offices that are planning to grow, staff will take on multiple roles in the early days. A small team should have a diverse set of skills because it is difficult, if not impossible, to find all the skills needed in one person, or in one type of person. Ideally, the team includes a range of complementary skills and backgrounds. A CDO is the leader who can create an environment where separate disciplines work in harmony toward the same mission.

The table below describes the range of tasks that can be part of creating a data-driven culture in government. These roles may be distributed across parts of government, such as in a performance office, an innovation office, or on the CDO team. The largest CDO teams incorporate many of the tasks below, while the smallest CDO offices may have one or two. The mix will vary based on the CDO mission. 

Tasks and Typical Duties for Staff in a Government CDO Organization

Task Type

Typical Duties

Business Process Analysis

  • Study the problem to be solved and identify the policy result to be achieved, then clearly articulate the business process challenge to the department responsible for results.
  • Clearly articulate the goals of the analytics project and how it will create benefit.
  • Interact with department end users of analytics projects (many teams call this the “client-facing” role).

Data Analysis

  • Clean and normalize large datasets.
  • Perform analytics to identify trends and underlying truths within a dataset.
  • Create business intelligence reports.

Data Visualization

  • Use GIS and other platforms to create spatial and temporal maps of policy issues such that insights can be seen visually.

Data Modeling, Data Engineering, and Data Science

  • Build and configure data infrastructure to facilitate analytics while ensuring accuracy, security, and reliability.
  • Clean and normalize large datasets.
  • Explore large datasets to look for patterns, trends and insights.
  • Develop sophisticated models applying decision science and/or machine learning to data problems.
  • Test models and refine based on test results.
  • Working with departments and other CDO staff, develop methods to operationalize models.

Performance Analysis

  • For CDO offices with performance management responsibility, work with departments on tracking operational data for stat meetings.

Project Management

  • For CDO offices with sufficient staff to have a dedicated project manager, keep large or long-term projects on track, and coordinate the efforts among the various skilled resources. In CDO offices without dedicated project management staff, the CDO can establish project management protocols and tools for the team.
Finding, Training, and Retaining Talent


The challenge of finding and hiring the right staff is significant, with many CDOs reporting that they have problems both finding the right staff and keeping them. Because data analytics in government is a new field, there isn’t an established career path or a universal training ground. There are a handful of graduate programs to train students specifically for jobs in data science or analytics in the public sector. The University of Chicago Harris School has developed a Masters-level program to train students for public service data science roles and graduated its first class in 2016. New York University’s Center for Urban Science and Progress also trains students to use analytics to address urban problems. Increasingly, analytics skills are being added to graduate program curricula, but it takes time for the new programs to produce results. Some government CDO offices hire graduates of public policy programs and train them in data analytics on the job.

Leveraging Partnerships to Create More Capacity

Many successful city CDOs extend their capacity with partnerships. For example, Pittsburgh CDO Laura Meixell has extended her team’s capacity by partnering with nearby Allegheny County and the University of Pittsburgh. This collaboration resulted in a regional data center, which hosts a combined open data portal and links their open data to that of other public sector agencies, academic institutions, and nonprofit organizations.

Many CDOs partner with local educational institutions for fellows, PhD projects, and interns. Boston runs a successful summer fellow program that includes both graduate students and college interns who have worked on substantive projects, including a homelessness data warehouse. Taking on summer interns and fellows does take an investment of time, but it can have long-term value when the work product is good, and if used as a process to recruit and screen future full-time employees.

Some CDOs have received pro bono or steeply discounted help along the way. Los Angeles CDO Lilian Coral credits the team at GIS mapping partner Esri with delivering far more than expected when they helped her build the GeoHub. New Orleans Analytics Director Oliver Wise has had outside help from UPS, Louisiana State University, Tulane University, and from a private-sector analytics firm.

Chicago CDO Tom Schenk has made extensive use of pro bono partnerships, with nearly $1 million in in-kind support from partners, including:

  • Allstate provided analytics support to a model predicting risk for elevators to prioritize those most needing inspection, work predicting which restaurants were most likely to fail their inspections, and a model to predict weather-related tree limb loss.
  • Sagence Consulting provided analytics support for a fraud detection model used by the city.
  • Other notable examples include a local web development firm that contributed support for user interface work; a firm that helped develop a model to predict business failures; a firm that explored a pothole response time optimization model; and a firm that contributed to data infrastructure for back-end data storage for analytics.

In addition, Chicago has made great use of civic tech volunteer help. The city’s Chi Hack Night group contributed more than 200 volunteer hours to a project last year for a bacteria-detection beach closure model. Volunteers also helped test the user interface of their OpenGrid open-source analytics sharing platform.  

In forming external partnerships, CDOs report that it is important to have a clear request. Just asking for help is not nearly as easy for an interested local corporate executive to respond to as a request for a specified number of hours of analytics work on a specific dataset or model.


Data Stewardship and Analytics

The day-to-day work of moving a government toward greater ability to use data consists of data stewardship and data analytics. No two CDOs are alike, and not all have responsibility for both stewardship and analytics, but the two functions are interdependent and are addressed together here.

Data stewardship refers to the data infrastructure to manage data, and the data governance that guides data collection and quality, such as in an open data program. Data stewardship is perhaps the least discussed and yet one of the key inputs to a CDO’s success, as it enables the analytics projects that build excitement and bring press attention.

The term analytics is used widely and with many definitions. According to the Business Dictionary, analytics “involves studying past historical data to research potential trends, to analyze the effects of certain decisions or events, or to evaluate the performance of a given tool or scenario.” For purposes of this paper, analytics refers to the work of turning data into insight, whether through descriptive reports and mapping, or via predictive modeling.

CDO responsibilities vary widely, with some focused mostly on data stewardship and their open data programs, and others more focused on analytics. While not all CDOs have jurisdiction over data governance, all CDOs have an interest in data quality and availability.

Improving data quality and consistency allows it to be shared with the public in an open data program and also enables data analytics projects. Even with high-quality data, there can still be a great deal of work in cleaning and standardizing the data so that it can be used in analytics projects. Some CDOs say they spend two-thirds of their time getting data in a format that can be manipulated. Others estimate that as much as 95% of the effort on an analytics project is to devoted to cleaning and organizing the data before it can be analyzed.

As shown below, a data analytics capability relies on a foundation of data infrastructure and data governance, the data stewardship tasks.


A foundation is laid with data governance, establishing the policies and standards by which government data is collected and managed, as well as the strategy for data collection and sharing (the what, why, and when). Upon that foundation, data infrastructure provides the mechanism for data to be collected and shared with the public on open data portals, as well as with other departments of government (the how). The result is both open data and internal data that can be used for mapping and descriptive or predictive analytics projects. The table below describes how these functions relate to one another.

                                              Operational Framework for Enabling Data-Driven Government


Questions Addressed


Data Governance

(Policy and Strategy)

Answers why, when, and what data is collected, generated, and shared within government and to the public

Setting standards for data quality, standards, and interoperability and determining how open data will be handled sets a solid foundation for data-driven decision making in government. Policies and standards should address data quality, consistency, frequency of updates, security, and accessibility. As discussed elsewhere, opening greater volumes of data can accelerate improvements in data quality. Strong data governance fosters interoperability and reuse of data across the enterprise. Data governance also addresses cross-agency sharing of who collects what, and for what purpose, to avoid duplication of data collection in different or competing formats.

Data Infrastructure (Services and Operations)

Answers how and where data is stored

Establishing the architecture of how data is stored and integrated across government can enable data-driven government by making data accessible. A range of formats are possible, from data warehouses to data lakes and data marts; regardless of the type, there must be a way to store and manipulate large volumes of data for data-driven government. Whether in traditional data centers or in the cloud, these systems must be monitored and maintained. Further, the provisioning of data in data warehouses enables more efficient data cleaning and normalization, steps that are critical inputs to any analytics project. Creating open data platforms allows data sharing with the public and across departments. Standardizing the processes internal to the CDO team for project management and execution further supports high-quality data infrastructure across the enterprise, and improves repeatability of solutions.

Data Analytics

Asks what is, why, and what if questions about the data

With large volumes of data available and with the tools for analytics, CDOs can look for patterns in their data and can gather insights that will improve service delivery to the public. Whether by mapping or developing descriptive statistics or predictive models, CDOs can advance the use of data to drive decisions by policymakers.


A good case study in the importance of data governance comes from Louisville, which hired its first Chief Data Officer in 2016. One of the motivating factors was to have someone responsible for data governance and enterprise-wide data management strategy. According to Theresa Reno-Weber, Louisville Chief of Performance & Technology, the initial open data work “really showed us the challenge of having multiple chefs in the kitchen – there needs to be one person who is really thinking strategically about data governance and using data as an asset both for us internally and for our public.”

The city has dramatically accelerated the use of data for decision-making in the last five years through the work of its Office of Performance Improvement and Innovation (OPI2) and its performance program, LouieStat. The city’s open data program has opened new channels to share data across government and with the public. In this process it became clear that the city needed a single point of responsibility for setting rules, policy and governance standards for the open data portal. Someone was needed to provide leadership and guidance to departments on what to release, when to release it, and how to release it, and to be the central point of contact for all of the data quality and data standards issues that arose as departments began sharing data. The new CDO will provide guidance to departments, to the IT organization responsible for hosting open data, and to the public.


Messaging to and Engaging Departments

Once a vision and roadmap and a team are in place, the CDO can begin engaging with departments. At this point, a written summary of the CDO’s charter is helpful as it will help make clear to departments what the team will and will not do. A one-page summary of the types of projects a CDO office will take on or an FAQ document might be helpful.

Engaging with Departments Across Government

A CDO can interact with other parts of government in a variety of ways. How a CDO chooses to engage with departments will depend on a variety of factors including how ready the departments are to make use of their data and how open the department heads are to the involvement of an outsider in their work. The type of engagement will vary from developing data dashboards to creating sophisticated predictive analytics models. The level of engagement will vary by department, but the overall philosophy of engagement can be established by the CDO.

In the private sector, where CDOs have been in place for longer than in the public sector, experts have studied the engagement of CDOs and have described two basic models — centralized and decentralized.[3]

In cities, the centralized model retains analytics resources in a single team under the CDO’s leadership and provides services to departments, functioning as an internal consulting firm in support of the analytics needs of government departments. Some departments will have their own analytics resources, typically the better resourced or statistics-driven departments such as police, but the majority will rely on the centralized CDO team.

A centralized pool of analytics talent allows sharing of specialized skills across the enterprise from a common hub. Given that highly-trained analytics staff can be expensive for government, this is a judicious use of resources. Another benefit of a central pool of talent is the efficiencies gained via peer support and collaboration among team members. A centralized team is better able to standardize tools and processes across government, which can save time and money and help develop deeper expertise in the chosen methods. The team can also facilitate cross-departmental data initiatives due to its citywide view of available data. A downside to the centralized approach is that it slows the growth of sustainable analytics talent in the departments. Further, it can be difficult to achieve scale with a small, centralized team.

The decentralized model creates distributed capacity across government by embedding talent in departments. In the decentralized model, each department is responsible for developing its own analytics capability, which might range in expertise from budget and policy analysts who can complete basic descriptive statistics to analysts with the skill to perform data science tasks such as predictive analytics. One advantage to the decentralized model is that analysts develop subject matter expertise that makes them valuable to their departments. Departments have more control over their analytics resources.

Putting decision-making and control of analytics in the hands of department heads leads to uneven attention and results across departments, with some investing heavily and others giving it low priority. In government, a drawback to this approach is that unless compelled to do so, many departments may not appoint an analytics officer at all. For data-focused employees, a decentralized model has the downside of not working with peers, and having a more limited career ladder.

While the centralized and decentralized models present opposite ends of a spectrum, some organizations employ a blended or hybrid model that uses the best elements of the two basic models. The hybrid model has been found to be the most effective in the private sector.[4] In government, most CDO offices follow the hybrid model.

Boston is a good example of a city that uses this model, neither fully centralized nor fully decentralized. The Citywide Analytics Team includes a variety of roles, from data scientists, data engineers, and GIS specialists to performance analysts. For some city departments, the team functions in a decentralized mode, supporting staff located in the department. Some, like the police department, have their own data analysts who can support day-to-day operations. For some departments, the team functions in a centralized manner, doing analytics work directly. For these efforts, the analytics team builds goodwill and relationships of trust with the department by solving tactical problems, which often means building a dashboard or map to meet a specific request of the department head.

One example is the Boston Fire Department, which relied heavily on the Citywide Analytics Team for help implementing a process to monitor how firefighters swap shifts – something that came under scrutiny after a series of negative reports in the press. At the time, the department didn’t have the recordkeeping capacity, technology, tools, or analytic capability to make sure it was following the rules. The centralized analytics team stepped in and helped with this concrete task, which met an immediate need. This built their credibility and created a spirit of partnership, which has carried over to other analytics projects, such as the recent project that alerts firefighters to building hazards when they are en route to a call, bringing together seven separate datasets from across the city into a single visualization.

Identifying Analytics Projects from Departments

For a CDO, coming up with a structured way of engaging departments in analytics projects can be challenging. Many government CDOs are new to their jobs and have not yet developed their strategy for engaging with and soliciting project ideas from departments. Interesting models for gathering input from departments for possible analytics projects are found in Chicago, New Orleans, and New York City.

The City of Chicago received a grant to create an open-source predictive analytics platform, the SmartData Platform, as a winner of the Bloomberg Philanthropies Mayors Challenge. To generate predictive analytics ideas, or use cases, for the SmartData platform, Chicago developed a methodology to systematically reach out to agencies across the city to gather input and to identify the most powerful ways to use predictive analytics to improve city services to the public. The process includes the following steps:

Chicago’s Use Case Identification Process



Commissioner Interviews

The data team held one-on-one meetings with Commissioners for 10 city agencies to introduce the SmartData Platform and explain the potential for predictive analytics to make department operations more effective. The meetings built high-level stakeholder buy-in and identified 32 possible use cases for predictive analytics.

Use Case Development Workshops

Four cross-agency workshops invited brainstorming on ways to use predictive analytics to improve city response to public needs. Departments were clustered by topic area and then asked to come up with ideas that spanned their “silos” of operations. The facilitated cross-agency sessions identified 108 unique ideas that spanned a wide range of topics.

Follow-up Commissioner Meetings

In follow-up meetings, each of the Commissioners reviewed the possible use cases and weighed in on which were highest priority in making meaningful operational improvements. These meetings also filtered out ideas better suited to tactical fixes rather than big data analytics. A set of seven evaluation criteria were used to rank each idea, ranging from data availability to level of impact on the public. 20 ideas remained based on these filters.

Final Priority-Setting by the Mayor

The ultimate decision for which use cases were addressed and in which order factored in mayoral priorities and the most pressing civic needs of the people of the City of Chicago. The final filter for the ideas ranked them according to the existing evaluation criteria as well as the urgency of the issue and degree of public value that could be generated. With this final filter in place, the team had seven use cases to work on over the two-year period of the grant.

Now, nearly two years after establishing this list of top-priority use cases, Chicago is sufficiently confident in this process that they aim to repeat it soon.

New Orleans has recently created an annual process to identify analytics use cases. Beginning in summer 2016, the NOLAlytics team, led by the CDO, developed a timeline for selecting analytics projects – they would solicit ideas in the fall, decide in early winter, and then have the year to execute on projects.

To generate interest and to advertise the availability of their analytics services to departments, they created a presentation to explain the power of data and the types of use cases they could take on. They shared this broadly across departments, with all senior managers, data analysts, and data coordinators across government.

They then hosted an open house to invite anyone interested in learning more to talk directly to the CDO and his team about possible analytics projects. The open house was valuable because departments that were not among the “usual suspects” for analytics work attended and developed ideas for new ways to use data to improve operations. The analytics team created a simple application form departments can use to submit possible analytics projects, and a checklist that departments can use to assess whether their project is ready to be addressed via analytics. As 2016 is the first year of the new process, it will take time before the city can determine how well this works.


New York City has another model of engagement, where the Mayor’s Office of Data Analytics (MODA) serves as an internal analytics consulting firm for the city. Owing to its reputation for delivering results, MODA receives a steady flow of analytics requests. MODA does not have a formal channel for project intake, but instead takes a “no wrong door” approach.

The Chief Analytics Officer (Director of MODA) is responsible for liaising with leaders in the Mayor’s Office and across the city to understand how analytics can help their top-line priorities. Other members of the MODA team develop relationships with more mid-level personnel across the city who have more tactical insight because their daily work is closer to the actual data.

Once an idea is generated, projects are prioritized based on the following criteria:

  • Potential for operational impact: How much of a difference is the project going to make? Will the change have a significant impact on operations? Is the agency clear about what it is looking for and how it will use the results?
  • Sponsorship: How committed is the executive sponsor at the agency? How many resources are being committed to the project by the agency? How many resources in addition to MODA are on board? Will there be enough resources to make the project successful?
  • Fit with MODA skills: Will the unique skills and tools that MODA has to offer provide leverage to make a difference in agency operations?

MODA only takes on projects that meet some or all of these criteria. This screening method improves organizational buy-in for projects, which increases the chances that analytical models built by MODA analysts will make it to an operational pilot stage, and eventually be integrated sustainably into an agency’s operational workflows.

There are many ways to engage with departments, but one lesson that came up repeatedly was the value of relationship building and deep listening.


Fostering a Culture of Data-Driven Government

The impact of a CDO can be significant and lasting – they have the power to foster a culture of data use across government and create distributed capacity. In moving their government toward data-driven decision-making, the CDO faces the classic tradeoff between doing the work for departments or “teaching them to fish.” In many cases, the CDO has no choice but to build distributed capacity.

Tyler Kleykamp, CDO for the State of Connecticut, is not alone among CDOs in having a staff of one. The executive order that created his office also created data coordinators in each executive-branch agency. His engagement method is to provide the platform and tools along with training so that the data coordinators can publish and use data within their department as well as across government. And he’s having an impact – by facilitating the release of both state public health drug overdose data and local crime data for drug offenses, he has enabled analytics both across departments and outside state government. Mapping has shown that overdoses might not occur at the same addresses as drug sales, but they are typically close by. Public use of and feedback about the data identified irregularities in the data that have resulted in quality improvements, and fueled a week-long series on addiction by a local interest group. The impact of this open data is not just local, with users as far away as an opioid overdose hackathon event in D.C.

Santa Monica, CA has made major advances in achieving a “culture of data” even before hiring their first CDO, which the city plans to do in early 2017. The city’s data-intensive Wellbeing Index has already created important shifts in culture and use of data. The project, which won the 2013 Bloomberg Philanthropies Mayors Challenge, aims to measure the degree to which people are thriving, and allow government to foster community and individual wellbeing. The Wellbeing Index team has held workshops on how to use data to inform policy, and has created open incubator sessions where teams are armed with data from the Wellbeing Index and asked to devise solutions to improve wellbeing gaps identified by the data. When the city hires its first CDO, the momentum on creating a culture of data-driven decision-making will already be well underway.

One of the best examples of creating culture change and building capacity across government comes from San Francisco, where the CDO’s office in partnership with the Controller’s office offers training to city and county staff on a variety of data skills through the SF Data Academy. Their goal is to allow city staff to “explore, refine, and enhance their skills in data analysis and visualization.” A variety of courses are offered on a regular basis, and customized courses can also be delivered to a department.

Courses include both classroom-based learning and online courses, designed to be accessible to a broad audience. Topics include skills such as business process mapping and specific tools such as Excel and Tableau. In addition to offering instructor-led classes, the SF Data Academy curated a list of online content covering a wide range of skills and tools that city staff can access. This model is now being emulated in Connecticut and San Diego, and has already been adapted and deployed by New Orleans.

The Pittsburgh CDO has responsibility for the city’s Lean Six Sigma process improvement project and is building capacity not just for data-driven government, but for innovation in general. The CDO team is active in supporting innovation across the city, including playing a part in the city's application for the U.S. Department of Transportation's Smart Cities Challenge grant competition.  

Chicago builds citywide capacity by connecting to its broad civic tech community, which consists of thousands of data and policy enthusiasts who volunteer to advance the public good with technology. Its CDO participates in several regularly-held civic tech meet-ups and events, including Chi Hack Night, one of the city’s largest civic tech volunteer groups. To spread data culture within city government, the CDO also holds training sessions on analytics, and has created a tutorial video on how to use OpenGrid, the city’s new open-source analytics and data visualization tool.

In Los Angeles, the CDO held open workshops for public servants in city departments to help them learn to use the new GeoHub data visualization tool. The series of workshops kicked off with one session that included representatives from over 20 city agencies. The goal of the workshops is to increase skills in data analysis and visualization, as well as to inspire more inter-agency conversation about the value of data in government operations. Los Angeles recently formalized its inventory efforts by convening 55 data coordinators from over 30 departments to provide uniform data standards and improve the city’s open data libraries on the GeoHub and its open data portal.

In New York City, the Mayor’s Office of Data Analytics partnered with the Department of Citywide Administrative Services (DCAS) to develop courses that train employees across the city on data analytics tools. They also developed a course for managers on how to lead their units in a more data-driven way.

Hiring a CDO offers a unique opportunity to advance data capacity across the enterprise. Regardless of the model they choose, a CDO has to build capacity in departments to have an impact across government.


CDOs in government are in a position to make a lasting difference in the lives of the public. One concrete example is reducing fire deaths. CDOs in Boston, New Orleans, and New York City have saved lives that might have been lost to fires.

  • In Boston, firefighters are equipped with data-powered protection from hazards, with a fully-integrated dashboard map of all hazard data in the city keeping them protected while en route to fight fires. Begun as a project of a fire dispatcher, the Citywide Analytics Team developed a visualization tool that integrates location-based information across seven databases showing where fire hydrants, biological and industrial hazards, and code violations are so that firefighters make it quickly and safely to their destinations.
  • New Orleans gets people out of burning buildings with strategic placement of smoke alarms to save lives, particularly among vulnerable populations. In response to a tragic loss of life, the fire chief asked the CDO for help prioritizing where to install smoke alarms in its free distribution program. Using varied data from city and national sources, the CDO team identified homes most at risk of fire and least likely to have smoke alarms. The result – in just one recent example, 11 people, including a baby, were saved from a burning building because of the alarm they got through this data-driven program.
  • NYC building inspectors prevent fires by prioritizing inspections at the riskiest addresses. The Mayor’s Office of Data Analytics used address level data to identify buildings that were a high risk for illegal conversion from single family into subdivided multi-family units, which is a fire hazard as families cook on hotplates instead of in kitchens, and may not have appropriate forms of egress. Using compliance data from the city’s 311 system, property tax payment data from the city’s Department of Finance, and foreclosure data from the Office of Court Administration, the Department of Buildings was able to get building inspectors to the highest-risk properties first.

With these and other visible successes, CDOs are becoming culture change agents in moving their governments toward data-driven decision making.

CDOs can also be pioneers in breaking down silos – for example, the work of the Los Angeles CDO to create the GeoHub inspired cross-departmental sharing of road closure data; now, fire trucks get to their destinations faster by avoiding roads that are closed due to construction.

CDOs can be data evangelists – by going to conferences, hackathons, and other events, CDOs can bring more attention to the data resources available to the public and can improve quality by shining the light on government data and analytics work.

Analytics officers at private-sector companies make money by analyzing our data and presenting it to us in ways that make it easy for us to spend money with just one click. A government CDO can do something much more powerful than suggest a purchase – they can analyze the data and present it in a way that lets a government decision-maker see new insights and make better decisions, creating greater public value. In government, CDOs are powering decisions about resource allocation to improve pedestrian safety, create more responsive and efficient transit routes, reduce opioid overdoses, and more accurately predict when beaches should be closed due to unsafe water.  

Today, data is plentiful but insight is far less common. The volume of data produced by state and local government is significant – Chicago alone produces seven million rows of data every day, from police reports to information about schools, libraries, and public parks. Analytics gives us a way to sift through the data and find insight and knowledge that can leads us to action. There is ample opportunity to use data for the common good – what an exciting time to be a CDO in government.

Profiles of Selected CAN Members

As has been described in the preceding pages, there is no one single way to structure an urban analytics organization, nor is there one common set of duties or priorities across organizations. To demonstrate the range of responsibilities and achievements of CDOs, brief profiles of selected analytics teams were created. Click here to see the profiles.


News Articles, Presentations, Research Reports, and Scholarly Journal Articles:

Brad Brown, David Court, and Tim McGuire, March 2014, “Views from the front lines of the data-analytics revolution,” McKinsey Quarterly.

Thomas H. Davenport, December 2013, “Analytics 3.0,” Harvard Business Review.

Mario Faria and Debra Logan, January 2016, “Staffing the office of the CDO,” Gartner.

Mario Faria, November 2015, “Successful organizational design principles for the office of the Chief Data Officer,” Gartner.

IBM Institute for Business Value, 2011, “The power of analytics for public sector: Building analytics competency to accelerate outcomes.”

IBM Center for the Business of Government, 2015, “Using innovation and technology to improve city services.”

Jascha Franklin-Hodge, Chief Information Officer, City of Boston, 2015, “Citywide analytics team, 2015 year in review.”

Julio Hernandez, Robert Berkey, Rahul Bhattacharya and Chad Vaske, March 2014, “How to become an analytics-driven consumer packaged goods company,” Accenture.

Julio Hernandez, Robert Berkey, and Rahul Bhattacharya, 2013, “Building an analytics driven organization: organizing, governing, sourcing and growing analytics capabilities in CPG,” Accenture.

Valerie A. Logan, February 2016, “The life of a chief analytics officer: a high-wire balancing act,” Gartner.

Brian McCarthy, Robert Berkey, Chad Vaske, 2015, “Launching an insights driven transformation building and sustaining analytics capabilities across the enterprise,” Accenture.

McKinsey, April 2016, “The need to lead in data and analytics.”

Thomas W. Oestreich and Frank Buytendijk, July 2016, “Why you need to rethink your data and analytics roles now,” Gartner.

Partnership for Public Service, November 2013, “From data to decisions III: lessons from early analytics programs.”

PwC, 2016, “Data-driven: Big decisions in the intelligence age.”

PwC, February 2015, “Great expectations: The evolution of the chief data officer.”

Svetlana Sicular, August 2014, “Big data analytics failures and how to prevent them,” Gartner.

Jeb Stone, “Centralized vs decentralized analytics: All you need to know,” April 22, 2012.

Andrew Therriault, 2016, “The future of data science in the city of Boston,” Presentation by Andrew Therriault, Chief Data Officer, City of Boston, 2016 Boston Data Festival.


Joy Bonaguro, Chief Data Officer, City of San Francisco, interview by author, April 1, 2016.

Andrew Buss, Director of Innovation Management, Office of Innovation & Technology, City of Philadelphia, interview by author, August 31, 2016.

Ray Campbell, Executive Director, Center for Health Information and Analysis, Commonwealth of Massachusetts, interview by author, September 21, 2016.

David Edinger, Chief of Staff, Office of Mayor Michael B. Hancock, City and County of Denver, interview by author, September 19, 2016.

Jascha Franklin-Hodge, Chief Information Officer, City of Boston, interview by author, October 26, 2016.

Kelly X. Jin, Policy Advisor to the U.S. Chief Technology Officer, Office of Science and Technology Policy, Executive Office of the President, interview by author, September 15, 2016.

Kristina Johnson, Senior Consultant, Edward J. Collins, Jr. Center for Public Management, John W. McCormack Graduate School of Policy and Global Studies, University of Massachusetts Boston, interview by author, September 9, 2016.

Tyler Kleykamp, Chief Data Officer, State of Connecticut, interview by author, February 1, 2016.

Barney Krucoff, Chief Data Officer, Washington, D.C., interview by author, September 30, 2016.

Laura Meixell, Analytics and Strategy Manager, Department of Innovation and Performance, City of Pittsburgh, interview by author, March 10, 2016.

Kevin Miller and Erica Garaffo, Data Analytics Team, City of San Jose, interview by author, May 25, 2016.

Lynn Overmann, Senior Advisor to the U.S. Chief Technology Officer, Office of Science and Technology Policy, Executive Office of the President, interview by author, September 15, 2016.

Theresa Reno-Weber, Chief of Performance & Technology, Louisville, Kentucky, interview by author, August 22, 2016.

Benjamin Shaffer, Director of Performance Management for the State of Rhode Island, interview by author, August 8, 2016.

Owen Stone, Senior Associate, Public Sector Innovation at Living Cities, interview by author, August 29, 2016.

Andrew Therriault, Chief Data Officer, City of Boston, interview by author, September 26, 2016.

Oliver Wise, Director, Office of Performance and Accountability, City of New Orleans, interview by author, September 19, 2016.

About the Civic Analytics Network

Based at the Ash Center for Democratic Governance and Innovation at the Harvard Kennedy School and funded by the Laura and John Arnold Foundation, the Civic Analytics Network is an affiliation of chief data officers from the largest and most innovative municipalities in United States. They are open data stewards, internal consultants, and performance managers. The network seeks to advance the use of data and analytics in municipal governance through facilitation of in-person meetings among members and production of research and documented best practices.

About the Ash Center

The Roy and Lila Ash Center for Democratic Governance and Innovation advances excellence and innovation in governance and public policy through research, education, and public discussion. The Ford Foundation is a founding donor of the Center. Three major programs support our mission: the Program on Democratic Governance, the Innovations in Government Program; and the Rajawali Foundation Institute for Asia.

This paper is copyrighted by the author. It cannot be reproduced or reused without permission.

Download PDF