This chapter begins with a definition and taxonomy of data, then analyses the implications of data use at each stage of the government data value cycle. It demonstrates how local governments can leverage data to generate public value and improve residents’ well‑being, from anticipating future trends and risks, to improving public services and supporting evidence-based governance through evaluation and impact assessment. The chapter also presents a tailored OECD data governance framework for the local public sector. The framework aims to help local governments target data governance elements at the strategic and tactical layers to generate public value and transition to a data-driven organisation. The chapter concludes with patterns and findings on data use and well‑being outcomes in cities.
Innovation and Data Use in Cities
3. Data use as a feature of policy making in cities
Abstract
The role of data in supporting local governments
Local governments increasingly recognise the roles of data in improving residents’ well-being. In their daily operations, local governments generate and assemble a massive quantity of data with diverse forms and characteristics. The insights derived from such data can contribute to the processes of knowledge creation, innovation and policy making. Indeed, enhanced access to and strategic use of public sector data can “lead to important value creation from economic, social, and good governance perspectives” (OECD, 2015[1]). The phenomenon is accelerated by socio-economic and technological trends such as the growing capacity for data generation, collection and dissemination, the recent developments in data analytics, and the emergence of a paradigm shift in knowledge creation and decision making. Despite these advances in capabilities and awareness, the full potential of a data-driven public sector still eludes many governments, failing to consistently result in tangible outcomes. Recognising data as a resource is the first step in transitioning toward building a data-driven government.
Definition and taxonomy of data
The fuzzy definition and interchangeable usage of the term “data” with related terms contribute to confusion. The term “data” has several meanings depending on the context and jurisdiction, and may refer to raw and unstructured data in various formats, personal information, administrative records, or factual information in various forms and formats. In some instances, “data” can be synonymous with “information”, while in others, the latter is specifically understood as “the meaning resulting from the interpretation of data” (OECD, 2015[1]). In this report, the term “data” covers only raw and unprocessed facts and statistics, while “information” represents the meanings conveyed through processed and assembled data. Subsequently, “knowledge” or “intelligence” are gained through the assimilation or internalisation of information. The distinction between these terms is crucial in understanding the cycle of data as well as the benefits it provides. For example, local governments may possess a large amount of data, yet have neither the tools nor the capabilities to clean, store, extract information, and generate public value from it. Likewise, governments may also suffer from “information overload”, making it difficult to leverage data to glean insights for decision making (OECD, 2015[1]). In this report, “government data” may be used synonymously with “public sector data”, which is generated, created, collected, processed, preserved, maintained, disseminated or funded by or for governments and public institutions at various levels, according to the OECD Recommendation of the Council for Enhanced Access and More Effective Use of Public Sector Information (OECD, 2008[2]). These data can be generated either directly or indirectly from public sector operations.
Box 3.1. Definitions of data-related terms
Data are understood as the representation of facts stored or transmitted as qualified or quantified symbols. In contrast to knowledge and information, data are assumed to have an “objective existence”, and they can be measured, namely in bits and bytes. Data can also be the result of datafication, a portmanteau for “data” and “quantification”, where a phenomenon or object is transformed into quantified symbols. Datafication should not be confused with digitisation, which refers to the process of encoding information into binary digits (i.e. bits) so it can be processed by computers. Data that have not been digitised cannot be processed by computers.
Information is often seen as the meaning resulting from the interpretation of facts as conveyed through data or other sources such as words. This meaning is reflected in the structure or organisation of the underlying source, including its hidden relationships and patterns of correlations, which can be revealed through data analytics. Information is therefore always context-dependent: it depends on the capacity to extract meaning from the information source; this capacity depending on available data analytic techniques and technologies as well as the skills and (pre-)knowledge of the data analyst.
Knowledge is understood as information and experience internalised or assimilated through a process, commonly referred to as “learning”. It provides the “learner” with the capacity to make effective decisions autonomously. Knowledge can be explicit, in which case it can be cost-effectively externalised to be communicated and embedded in tangible products, including books, standard procedures and intangible products such as patents, design and software. But it can also be tacit, based on an “amalgam of information and experience”, which is too costly to codify and thus to externalise.
Source: OECD (2015[1]), Data-Driven Innovation: Big Data for Growth and Well-Being, https://dx.doi.org/10.1787/9789264229358-en.
Understanding different types of data and their implications will help governments establish an effective framework for data governance, open the private sector to partnerships with the public sector, and most importantly equip residents (i.e. data generators) with better understanding of their rights and risks associated with data use. However, key actors in the data ecosystem, including local government agencies, the private sector (e.g. IT infrastructure companies, data service providers, data-dependent entrepreneurs, etc.) and residents do not seem to fully grasp the differences between data types and levels of access (i.e. openness). For example, not all data generated by the public sector is non-personal, which can be readily made available to the public. Similarly, data collected and produced by local governments is not necessarily public data until it has been uploaded and shared in the public domain (e.g. a city’s open data portals). On many occasions, the term “public” can be used to denote either ownership or domain.
The overlapping elements of different data types show that a taxonomy of data is needed to provide a common language for actors in the data ecosystem. Different data taxonomies have been proposed to provide key actors with a common language. One such categorisation comes from the Data Collaboratives (Verhulst, Young and Srinivasan, n.d.[3]) where (corporate) data is classified into four quadrants based on whether data is personally identifiable and whether it is disclosed by the subject or observed by a controller: (1) disclosed personal data, (2) observed personal data, (3) disclosed non-personal data and (4) observed non-personal data. For instance, “disclosed personal data” refers to “personally identifiable information actively and intentionally shared by an individual, entity or group for a specific reason”, including registration records. While “observed non-personal data” covers “information with no personally identifiable elements that is passively collected by an entity prior to any use” such as satellite and aerial imagery.
However, the binary dichotomy in this taxonomy may not fully capture the nuances of certain data types. As such, in an attempt to introduce a more comprehensive categorisation, the OECD (2019[4]) proposed a data taxonomy based on four main aspects: (1) personal data and the degrees of identifiability, (2) the domain of data, (3) the manner data originates and (4) possible access control mechanisms (See Box 3.2).
Box 3.2. OECD taxonomy for the governance of data access and sharing
Data comes from various sources and possesses different policy implications depending on its characteristics, such as the degree of anonymity, accessibility and openness, and how it is collected, stored, processed and used to generate value. Understanding different data types can help public sector organisations adopt a more differentiated and comprehensive approach to data use. The OECD (2019[4]) provides a taxonomy of data based on four main dimensions:
Personal data and the degrees of identifiability: Personal data is often referred to as “any information relating to an identified or identifiable individual (data subject)”, according to the OECD’s Recommendation of the Council Concerning Guidelines Governing the Protection of Privacy and Transborder Flows of Personal Data (2013[5]). However, the demarcation between personal vs. non-personal is increasingly blurred as developments in data analytics facilitate the process of linking non-personal data to identified or identifiable individuals (OECD, 2019[4]; Ohm, 2009[6]; Narayanan and Shmatikov, 2006[7]). Instead of the traditional binary split, the ISO/IEC 19941 (2017[8]) elaborated five categories of data based on identifiability (the extent to which a given data can be traced to an identity). The data categories include: Identified, Pseudonymised, Unlinked Pseudonymised, Anonymised, and Aggregated.
Domain of data: Like personal vs. non-personal data, the dichotomy between public-sector vs. private-sector data does not capture nuanced cases where data are jointly collected and processed via public-private partnerships, or where data is generated by the private sector thanks to public funding. The distinction also creates confusion with public (domain) vs. private (proprietary) data, which refer more specifically to the openness of data. Lastly, the split does not account for personal data. Therefore, this aspect of the taxonomy proposes three non-mutually exclusive categories where data may be situated, namely: Personal, Public and Private domains. Personal domain refers to “all personal data relating to an identified or identifiable individual for which data subjects have privacy interests”. Private domain refers to “all proprietary data that are typically protected by intellectual property rights…or by other access and control rights”. Public domain data refers to readily accessible and free-of-charge data not protected by intellectual property rights. These categories demonstrate different roles various stakeholders may play in each stage of the data cycle, attesting to the complexity of establishing comprehensive data governance.
The manner data originates: The manner data originates focuses on the co-creation process, where various stakeholders collect and generate data. The categories proposed – Volunteered, Observed, Derived, and Acquired data – seek to illuminate not only the extent to which individual data subjects are aware of the data generation process, but also the degree of stakeholders’ involvement. While volunteered data is actively and intentionally provided by data subjects (e.g. social media profiles), observed data is passively generated by individuals (such as GPS locations or online activities/behaviours) and actively recorded by data controllers. Derived data is synonymous with data processed by analytics, with the data subject having little awareness of how data originates (credit scores based on financial history). Finally, acquired data is typically purchased from third parties with stringent restrictions regarding sharing and re-use.
Data-access control mechanisms: This covers a variety of technical mechanisms for assessing data such as ad-hoc downloads, Application Programming Interfaces (API), data sandboxes, etc.
Source: OECD (2019[4]), Enhancing Access to and Sharing of Data: Reconciling Risks and Benefits for Data Re-use across Societies, https://dx.doi.org/10.1787/276aaca8-en.
Implications for cities throughout the government data value cycle
Identifying the pathway and tasks needed to transform government data into information and knowledge can help local authorities navigate and maximise the value of data at each stage (OECD, 2019[9]). Therefore, van Ooijen, Ubaldi and Welby (2019[10]) proposed a government data value cycle (Figure 3.1), identifying four stages through which government data flow and the associated values generated during each stage: (1) data collection and generation, (2) data storage and processing, (3) data sharing, curating and publishing, and (4) data use and re-use.
This analytical framework, presented as a feedback loop of value creation, begins with the data collection, generation, storage and processing stages where impacts and implications are restricted to the internal administration. Only during the two final stages – data sharing and data use – does the wider public noticeably experience the value generated from government data. Further, the data, information, and insights generated from these two final stages return as inputs for the first stage, creating a feedback circuit for the government data value cycle. It should be noted that improper handling during any stage of the cycle might cascade into subsequent stages. This section discusses the implications, risks and barriers to data use that local governments might face at various stages of the government data value cycle.
Figure 3.1. The government data value cycle
Source: OECD (2019[9]), The Path to Becoming a Data-Driven Public Sector, https://dx.doi.org/10.1787/059814a7-en.
Data collection and generation
Data collection and generation lay the groundwork for data use. Data collected, generated or obtained by local governments comes from various sources. They can be personal data collected from residents as part of an administrative process, internal records obtained from government activities or customer data generated from public-private partnerships for service delivery. Local governments can also source data from other public or private open data portals. In recent years, a large share of Internet of Things (IoT) sensors embedded in urban infrastructure such as transport, energy, healthcare and social services have generated large amounts of data. Indeed, some data are readily available to municipal administrations, requiring little effort for data collection. Others might require a more coherent strategy to allow local governments to obtain the data.
It is undeniable that local governments generate a huge amount of data that allows them to effectively carry out their everyday functions. However, these data tend to be collected with a “single-use only” purpose. The lack of standards for data collection and this “single-use only” mindset can lead to collection of duplicate data, an extra burden on service users and increased costs of processing and hosting data.
Additionally, local governments need to ensure that their data efforts not collect unreliable and outdated data that would hamper the ability to glean valuable insights. In their efforts to move quickly toward a data-driven organisation, local governments risk generating data from non-representative or biased samples by not account for the digitally excluded population. This would pose a problem should cities aim to leverage such data for targeted public policy intervention.
Data storage and processing
Storing and processing data carries privacy and security implications that can either improve or undermine public trust in government. As mentioned, the lack of standards in data collection and coding makes integrating sources and linking up data a challenge for municipalities. Potential problems with data processing and storage stem from the lack of data management and analytics capacity among municipal staff, and the lack of a coherent strategy and data culture within the organisation.
Even though linking up data would allow municipalities to construct more comprehensive datasets for in-depth analysis, it also poses privacy risks in case of security breaches, especially when it comes to personally identifiable data or data that can easily be de-anonymised. Data linkage, when leading to a single repository or database, can create a vulnerability that could compromise the government data (including personal data collected from residents) on which many municipal services rely. Due to these concerns, in many jurisdictions “the separation of linkage and analysis processes is therefore considered as best practice for confidentiality” (OECD, 2019[4]). Different trusted parties thus undertake the tasks of linking up data and data analysis.
City governments face a challenge to ensure that municipal networks and digital infrastructure are secure as online activities proliferate and cities commit to digitalisation of public services. Since the beginning of the COVID-19 crisis, more people found their professional and personal lives relocated online. During the first lockdowns, the number of people working from home in France increased tenfold; online shopping activities doubled. According to France Digitale, the number of cyber-attacks increased fourfold, both in frequency and intensity (Midena, 2020[11]). Ensuring the security of municipal systems should be government’s priority at every step of their handling data. It is particularly crucial during the data storage and processing stages, where responsible agencies must deal with a wide range of generated data, some of which is personally identifiable or proprietary.
Box 3.3. Immunising municipal networks against digital risks
From Newark (New Jersey) to Baltimore (Maryland), American cities of all sizes in recent years have been highly susceptible targets of digital breaches. In March 2018, a ransomware cyberattack brought municipal networks in Atlanta, GA, United States to a standstill. The hackers held the city hostage for nearly a week, demanding USD 51 000 of payment in bitcoins. The security breach stopped public Wi-Fi and reduced many municipal services to pen and paper. The Atlanta Municipal Court resorted to manual processing of cases. Law enforcement officers went back to writing incident reports by hand. Meanwhile, residents were unable to pay their traffic tickets and water bills, or apply for and renew business licences. The city also temporarily lost a huge amount of data in the forms of online correspondence, police camera footage and legal documents. In the end, Atlanta refused to pay the ransom and spent more than USD 17 million to upgrade security systems. The city also implemented a cybersecurity framework, engaged in dialogues with federal and state agencies, and increased digital security awareness among municipal employees and residents.
Source: Freed (2019[12]), One year after Atlanta’s ransomware attack, the city says it’s transforming its technology, https://statescoop.com/one-year-after-atlantas-ransomware-attack-the-city-says-its-transforming-its-technology/; Sneed (2019[13]), What Cities Can Learn From Atlanta’s Cyberattack, https://www.bloomberg.com/news/articles/2019-10-29/what-cities-can-learn-from-atlanta-s-cyberattack
Data sharing, curating and publishing
Data sharing practices vary across public organisations, due mostly to the respective legal frameworks (or the lack thereof) in each country. At one end of the spectrum, countries such as Korea, Portugal, Israel and Estonia adopt a proactive approach to intra-government data sharing (OECD, 2019[4]; OECD, 2019[9]). Motivated by the “once-only principle” where public organisations should not collect the same data as another public-sector database, Estonia’s X-Road initiative enables information systems from various public and private organisations to securely link up and exchange data. Another example of the “once-only principle” implemented at the local level includes was launched by the city of Monheim am Rhein, Germany, where residents can access a variety of public services through a single entry of their personal information. Thus, residents are not required to provide the same data multiple times to access public services provided by different city agencies. The initiative aims to reduce administrative burdens for residents and businesses, and also minimise digital risks by building a single, secure and centralised repository of data in the city’s system (European Commission, 2019[14]). On the other end, countries such as the United Kingdom do not institutionalise a formal mechanism for data sharing between public bodies, implying that data sharing can occur on an ad-hoc basis or after one-off requests (OECD, 2019[4]).
Data sharing at the local level can depend on national legal and strategic frameworks. In their absence, local governments’ efforts to share and leverage data might become complicated. During the Expert Workshop on Boosting City Government Capacity to Innovate and Use Data for Better Policies and Resident Outcomes, organised by the OECD and Bloomberg Philanthropies, the Director of GobLab UAI, a public innovation lab in Chile, noted that local governments in Chile struggle to utilise administrative data, which is centralised at the national level despite most data collection occurring locally. To access these data, local administrations must enter into a data-sharing agreement with the national government, which can take up to a year and might hinder efforts to leverage data in an effective and timely manner. The country thus is considering a change of legislation to allow easier access and integration of administrative data at the national level. This example shows how systemic hurdles to data use cannot be resolved solely by local governments.
Publishing government data enables external stakeholders such as the private sector, concerned residents, and the civil society, to use, re-use and generate even more social and economic value beyond their intended use by the government. In recent years, there is a noticeable trend towards the open data movement where (in the absence of conflicting interests) government data are widely expected to be published in a user-friendly format online (see “Data openness”, below).
Data use and re-use
Data use and re-use represent the last stage of the government data value cycle, where the value derived from data are most discernible to the wider public. Local governments with a mature approach to data rely on data analytics to derive knowledge for their targeted interventions. Data analytics are “the process of crosschecking, cleaning, reorganising and modelling data for decision making” (Auditor General for Wales, 2018[15]). Analysis based on inaccurate statistical reasoning, erroneous modelling, and biased algorithms would undermine the effectiveness of intended interventions. While the effectiveness of data-driven public policy could suffer if undetected risks and defects at earlier stages in the cycle have detrimental effects on the intended interventions, the use and reuse of data is nonetheless the key to anticipating trends and risks, improving public service delivery, evaluating performance, assessing impacts and monitoring programme outcomes (see “Unlocking the value of data for the local public sector”, below).
Thus, data and the knowledge derived from data should serve more as a tool to solve public problems, rather than the solution in themselves. A data-driven project is only as useful as the issues that it attempts to resolve. Drawing from their experience teaching “Data Science project scoping” to public sector managers in Chile, GobLab UAI note that government agencies at both the local and national level had a hard time with problem definition, struggling to define the issues that their data-driven project or service is trying to solve. Unsound problem definition would result in data being used in a counterproductive manner, potentially distorting insights and leading to ineffective interventions. When data is finally used (and re-used), it can yield the most concrete impacts for governments and contribute to the improvement of residents’ well-being.
Unlocking the value of data for the local public sector
Local governments have a critical role in leveraging data to improve residents’ well-being. Municipalities can use data to adopt a more customer-oriented approach, ensuring that public sector activity is motivated by tangible outcomes for residents rather than by symbolic motives like ideology or media attention. Additionally, municipalities can use data to create feedback loops that build equity and capture emerging trends to forecast future needs. Indeed, data enables governments to foresee trends and risks, allowing them to minimise or even pre-empt potential crises/problems. Secondly, local governments can leverage data to guide public service design and delivery in a more evidence-based, practical and efficient manner. Lastly and most importantly, data allow local governments to learn from their successes and shortcomings by measuring outcomes and iterating their interventions. Such data-driven self-scrutiny can help public agencies mature from making isolated, project-based investments to undertaking sustainable long-term, higher-impact initiatives (Salge and Vera, 2012[16]).
Data-driven monitoring and assessment also link to the previously mentioned improved service delivery and anticipation. While a focus on users and learning from implementation may seem like obvious goals for local governments, impact assessment and outcomes measurement are also key to building an accountable and transparent public sector.
Anticipating future trends and risks
Governments can leverage data, including historical and real-time data on geographical location, sensors and evidence from previous interventions to detect emerging societal, economic or natural developments, forecast future risks and formulate targeted policy interventions (see Box 3.4). Through integration of predictive analytics into public service planning and delivery, municipalities can develop insights into community challenges and residents’ unmet needs. Different from descriptive statistics, predictive models leverage large datasets of current and historical observations to predict patterns and forecast outcomes. When combined with real- and near-time data feeds, predictive analytics provide local governments with powerful, actionable insights to undertake preventative and pre-emptive measures to deliver more effective and timely interventions. In practical terms, predictive analytics allow cities to “focus the allocation of scarce resources, identify adverse events, and ascertain the effectiveness of tested interventions” to deliver better outcomes for residents (Toderas and Manning, 2019[17]).
Recent developments in data analytics including the ease, volume and rate at which data are collected and stored, and access to more powerful software and hardware that enable local governments to improve their ability to conduct predictive modelling. Despite gaining traction among local governments, predictive analytics generally remains at a nascent state. Indeed, while the private sector long integrated predictive analytics in various domains (e.g. marketing, telecommunication, insurance and retail, etc.), the adoption of such methods among public sector organisations leaves much to be desired. The UK Local Government Association (2020[18]) found that predictive analytics by local governments tends to be experimental or exploratory, small-scale and limited in scope. Projects employing predictive analytics tend to be confined to areas such as child protection, personal social care, housing repair, credit control and inquiry handling. The research identified a set of interlinked factors that influence the adoption of predictive analytics in local governments, including but not limited to (UK Local Government Association, 2020[18]):
corporate understanding of the value and potential of predictive analytics
enabling environments for digital innovation
strong data expertise and/or ability to acquire such expertise
support from frontline staff whose services leverage predictive analytics
public confidence in data science and the potential of data-driven decision making
extreme budget constraints that force municipal governments to prioritise their investment in the most targeted ways.
Box 3.4. Examples of predictive analytics for targeted interventions at the local level
East Sussex County Council, United Kingdom, partnered with the UK Behavioural Insights Team, using collision data collected over a decade to reduce dangerous road accidents. The team developed predictive models that "test historically held beliefs, predict future behaviours and recommend how interventions could be better targeted.” The team found that occupational drivers are not disproportionately involved in serious collisions. Most of these collisions involved local drivers or those who had previously been caught speeding.
Jakarta, Indonesia, collaborated with start-up company Qlue to predict and prepare for floods. Jakarta’s Smart City Unit and Qlue use historical data on water level, weather conditions and trends in residents’ complaints to predict the intensity of floods in specific locations for the following year. The partnership would soon expand to predict and prepare for dengue hotspots.
London Borough of Barking and Dagenham, United Kingdom, combines data from their ethnographic research with predictive analytics tools provided by private company Xantura to pre-emptively determine households at risk of homelessness. Case workers reach out to identified households for a consultation and to offer interventions. This proactive approach increased the success rate of homelessness prevention and led to better adoption of prevention services offered by the local authorities.
New Orleans, LA, United States, produced a predictive analytics model to identify households ill-equipped with smoke detectors and those most likely to suffer from fire fatalities. The project was a collaboration between the New Orleans Fire Department (NOFD) and Office of Performance and Accountability (OPA). As NOFD did not have historical data on locations of smoke detectors, OPA resorted to data from the U.S. Census Bureau’s American Housing Survey and the American Community Survey. The model was instrumental in helping NOFD identify households at risk and install 8 000 smoke detectors from 2014 to 2016.
In Scottsdale, AZ, United States, the Water Department uses predictive analytics to determine planned water use for their annual water order. The data analysis saves the department nearly USD 500 000 each year.
Source: OECD (2019[9]), The Path to Becoming a Data-Driven Public Sector, https://dx.doi.org/10.1787/059814a7-en; The Behavioural Insights Team (2017[19]), The Behavioural Insights Team Update Report 2016-17, https://www.bi.team/publications/the-behavioural-insights-team-update-report-2016-17/; Basu (2016[20]), Jakarta’s plans for predictive government, https://govinsider.asia/innovation/jakartas-plans-for-predictive-government/; Local Government Association (2020[18]), Using predictive analytics in local public services, https://www.local.gov.uk/using-predictive-analytics-local-public-services; Hillenbrand (2016[21]), Predicting Fire Risk: From New Orleans to a Nationwide Tool, https://datasmart.ash.harvard.edu/news/article/predicting-fire-risk-from-new-orleans-to-a-nationwide-tool-846; What Works Cities (WWC) Certification.
Improving local public service delivery
Cities, typically regarded as testbeds for innovation, have a unique position to leverage data for improved public service delivery. These data can be used to “inform and improve policy implementation, the responsiveness of governments, and the activity of providing public services” (OECD, 2019[9]). Indeed, cities produce, generate and store an enormous volume of valuable data. By 2022, it is generally expected that “cities will host at least 10 billion out of 14 billion data-generating devices in use in OECD countries” (OECD, 2015[1]; OECD, 2012[22]; OECD, 2010[23]). Additionally, a large share of the 65 million sensors embedded in infrastructure around the world are in urban settings across health care, environmental, transport, water and energy sectors, generating a huge swath of data. The sheer volume and variety of data across key policy sectors make cities an ideal environment for data-driven innovation.
The OECD/Bloomberg Philanthropies Survey on Innovation Capacity in Cities shows that, while cities produce and generate a large amount of data, its availability across policy sectors remains unbalanced. When asked whether surveyed cities possess sufficient data for decision making (within the context of innovation work), most report having data for transport and mobility (77%), land use (70%), and waste and sanitation (60%). Meanwhile, fewer than half report having data for areas such as health (48%), labour market (41%) and culture (37%). The availability (or unavailability) is due mostly to factors such as the nature of these policy sectors, the strategic priorities of municipal government and most importantly local governments’ capability to gather and collect relevant data.
Figure 3.2. Cities have varying degrees of data across policy areas
Note: Out of 147 participating cities for both OECD/Bloomberg Philanthropies surveys, 142 responded to the question “Does your city have sufficient or insufficient data in the following areas to advance your innovation work?” Surveyed cities were asked to select all policy areas that apply; every policy area chosen corresponds to one unit.
Source: OECD/Bloomberg Philanthropies (2018-20), Survey on Innovation Capacity in Cities.
Notwithstanding the unevenness in data availability, all this data holds potential for cities to enhance the efficiency and productivity of their urban systems, facilitate new business opportunities and improve urban governance (OECD, 2015[1]). As Table 3.1 conveys, the abundance of data flows collected both in near/real-time and in the past across key urban systems such as transport, land management, and water, waste and sanitation allows analysis at a remarkable depth and rate. The granularity of urban data flow is crucial for targeted intervention, enhancing the efficiency of urban systems.
The use, re-use and sharing of public sector data can generate products and services that “contribute in a variety of ways to improved efficiency and productivity within the public sector,” making data a driver “of growth, employment, as well as of improved public service delivery and more efficient, transparent and participatory governance.” (OECD, 2015[1]). Leveraging of data by cities can also “create societal benefits like less pollution, fewer traffic jams, improved tracking of disease outbreaks, greater energy efficiency, new agriculture services, novel applications to transform citizen experience interacting online with government, and lower costs,” all of which could lead to improved well-being for urban residents (Janssen et al., 2017[24]).
Table 3.1. Leveraging data to enhance the efficiency and quality of urban systems
Source: OECD/Bloomberg Philanthropies (2018-20), Survey on Innovation Capacity in Cities; City of New Orleans (n.d.[25]), Code Enforcement Abatement Tool: A NOLAlytics Project, https://www.nola.gov/performance-and-accountability/reports/nolalytics-reports/nolalytics-blight-abatement-tool-brief/; Results for America (2018[26]), How to Clean City Streets? Los Angeles Begins by Collecting New Data, https://results4america.org/wp-content/uploads/2018/12/LosAngelesCaseStudy_Final.pdf.
The public value of government data extends beyond local authorities holding and employing these data. Published datasets can complement privately held proprietary data, giving stakeholders a holistic perspective and improving their problem-solving capacity. When government data are readily available (either for free or for purchase), potential users of such data, in combination with data analytics, can increase the benefits of data re-use and further magnify the public value of data across society.
There is an economic case for a more data-driven and data-capable public sector, with studies estimating the direct impact (i.e. benefits to data providers), indirect impact (i.e. benefits to data users) and induced impact (i.e. benefits to the wider economy) of government data. According to a report from Deloitte for the UK Department of Business, Innovation and Skills, “direct economic impact (as revenue for public sector information holders) is estimated at around USD 130 million, while the indirect impact on data users and suppliers of data public sector information is between USD 1.6 billion to USD 2.4 billion annually. The wider indirect and induced impact of public sector information was conservatively estimated to be around USD 6.5 billion per year.” (Deloitte, 2013[27]). Other studies on public sector information re-use at the EU level estimated the market to be around USD 38 billion in 2010 (Vickery, 2011[28]; OECD, 2019[4]). Meanwhile, its aggregate indirect and induced economic impact across 27 European economies is estimated at USD 165 billion annually (equivalent to 1.5% of their GDP). Through greater sharing and use of data, cities enable residents and entrepreneurs to develop new urban solutions, stimulate competition and bring down the marginal costs of urban services (Cohen, Almirall and Chesbrough, 2016[29]).
Monitoring, evaluation and impact assessment
Successful monitoring, evaluation and impact assessment begins with a strong culture of evidence-based governance, where data is used, re-used and shared to monitor and evaluate social and economic programmes, support evidence-based decision making and increase public accountability. For example, the City of Moscow, Russia, uses healthcare service indicators to assess the performance of healthcare systems and assist a wide range of health-related research activities (Box 3.5). However, while there has long been discussion of how data can produce evidence-based evaluations of government activities and other public sector innovation, there is not yet a strong culture of evidence-based governance. As of April 2020, only 15 of the 153 local governments in the WWC Assessment programme had defined standards, methodologies or tools to help staff evaluate practices, programmes or policies (see Figure 3.3).
Figure 3.3. Few cities have standards, methodologies or tools for data-driven evaluation
Note: Out of 153 cities participating in the WWC Assessment programme as of April 2020, 15 cities meet the criterion for "Your local government has defined standards, methodologies, or tools to help staff rigorously evaluate practices, programs, and/or policies".
Source: WWC Certification Database.
Box 3.5. Integrated Medical Information and Analytical System in Moscow, Russia
The Integrated Medical Information and Analytical System (IMIAS) improves the quality of healthcare delivery in Moscow by centralising the electronic medical records of Muscovites. IMIAS facilitates access to healthcare services online – such as locating the nearest medical institutions, scheduling an appointment or accessing medical e-records – and reduces the administrative burden on medical personnel. By continuously updating non-sensitive data from patients in real time, the system provides authorities with performance metrics like the number of patients, waiting times, length of visits and estimated cost savings, which can be used to improve Moscow’s healthcare system.
Source: OECD/Bloomberg Philanthropies (2020), Survey on Innovation Capacity in Cities.
Evaluation contributes to learning processes (with the aim of improving, replicating or even abandoning a policy or activity) and informs policy makers about the effectiveness of interventions. Generally, evaluation refers to the methodical and objective assessment of past and present projects, programmes or initiatives. Such assessments can take place at various stages (e.g. project design, implementation, outputs). As understood by What Works Cities and their partner cities, the ability to measure, monitor and evaluate public services and programmes through data allows local governments and residents to determine whether policies, programmes and innovation pilots really produce positive outcomes and impacts relative to their objectives.
With empirical evidence, residents and local governments can have informed debates and decide which initiatives receive funding, staffing and support. Such data lead to more efficient budget allocation and improvements to project implementation, and they support comparative analyses that help address the “very fragmented and dispersed nature of public innovation” (Vries, Bekkers and Tummers, 2016[30]).
Understanding the progress and tangible outcomes (or a lack thereof) leads to a more refined process of implementation that yields more substantial results. Instead of blindly investing in any project, resources could be re-allocated to those that show signs of achieving the city’s strategic goals. From the WWC Assessment programme, 41% of cities indicated that they leverage the insights from data to align their budget process with their strategic goals (see Figure 3.4).
Figure 3.4. Cities that leverage data to align budget processes with strategic priorities
Note: Out of 153 cities participating in the WWC Assessment programme as of April 2020, 63 cities meet the criterion for "Your local government uses data to align its budget process with its strategic priorities".
Source: WWC Certification Database.
According to the 2020 OECD/Bloomberg Philanthropies Survey on Innovation Capacity, as part of the Dallas 365 Plan, the city of Dallas, TX, United States, publicly tracks its progress in 35 performance measures aligned to the city’s six strategic priorities. These indicators help the city’s Office of Budget benchmark performance and allocate resources during the budget development process. Likewise, the City Council of Bristol, United Kingdom, leverages data to align its budget process with its strategic priorities. Despite being distinct mechanisms, budget setting closely links to the Service Planning and Performance monitoring process. The City Council also produces an annual Business Plan with actions and objectives underpinned by data and evidence-based priorities. A more systematic data-driven approach to governance, like those in Dallas and Bristol, might move the public sector away from wasteful spending toward projects stronger in substance and sustainability (OECD, 2020[31]; OECD/Eurostat, 2018[32]). Ultimately, using quantitative metrics to guide decision making around governance could lead to more tangible outcomes and impact. With proper evidence, governments and the public would, for example, be able to discern whether a given public sector innovation reduced CO2 emissions, or if a given policy change led to greater inclusion of minority groups in economic growth over time (Gault, 2018[33]).
Box 3.6. Examples of city governments leveraging data to monitor, evaluate and assess the impact of their programmes
Adelaide, Australia, leverages data analytics to learn who uses citywide infrastructure (e.g. digital infrastructure for students and tourists, roads, parking and loading bays for SMEs, etc.) and how. Monitoring of physical infrastructure allows Adelaide to schedule maintenance and replacement projects with greater accuracy, optimising municipal resources such as their workforce and equipment.
Amsterdam, Netherlands, has a special IOS department tasked with measuring policy outcomes and producing relevant data across areas like health, safety and sustainability. Meanwhile, the Chief Technology Office assesses the achievement of their own innovation projects and initiatives. Amsterdam is currently exploring a City Innovation Index – a framework to measure innovation efforts in general throughout the city.
Barcelona, Spain, measures the number of people at risk of social exclusion who do not have access to public aid offered by the city. This allows the municipality to understand the needs of residents and provide them with better social care services.
Bologna, Italy, measures a range of indicators such as air quality (i.e. level of PM10) or short-term housing rentals to support social and economic programmes in the city.
Bristol, United Kingdom, through the Strategic Intelligence and Performance Team, provides evidence-based reporting on a range of high-level datasets to inform policy decisions. Examples include the Quality of Life survey that covers topics such as health, lifestyle, community, local services and public perception of living in Bristol. The final report contains 50 indicators and analysis of almost 6 000 comments about the changes residents want to see. Since June 2020, the Team also releases a new Strategic Intelligence bulletin with reports including Ward Profiles (containing datasets on population, health, education, crime and quality of life), Key Facts (major facts and infographics about life in Bristol), etc.
Through data analysis and resident feedback, the Department of Human Service Programs (DHSP) in Cambridge, MA, United States, identified that their preschool enrolment waitlist practices were not providing equitable opportunities for all families. To correct this, DHSP implemented a lottery system for the 2020-21 school year and increased the scholarship fund for qualifying families, allowing them to offer 40% of available seats to families with household incomes at or below 65% of the U.S. Department of Housing and Urban Development median income for Cambridge, even though this group only made up 29% of the lottery applicant pool.
Curridabat, Costa Rica, put in place a tailor-made monitoring system to assess the progress of its ecosystem services. The assessment is part of Curridabat’s Sweet City vision – a new model of urban planning to bring nature and biodiversity back to the city’s centre. Further, Curridabat tries to assess citizen engagement, to understand their needs and find better ways to communicate with residents. The city also designed and tested a survey on perceived happiness, which serves as one of the metrics to assess residents’ well-being.
Montgomery, AL, United States, introduced new software to identify and prevent blight, which draws insights from sources such as census data, utility data, building permits, housing code enforcements and 311 services. The programme enhances municipal staff’s capacity to single out vacant and abandoned properties, and shorten the process to remediate these properties.
Philadelphia, PA, United States, improved enrolment in the Philadelphia Senior Citizen Water Bill Discount Program by 15% with data-driven outreach. More than USD 125 000 went back into the pockets of eligible seniors.
Seattle, WA, United States, designed and executed a randomised control trial to test the effectiveness of communications aimed at tackling defaults on parking tickets and traffic camera citations. Data analysis showed that about 40% of parking tickets and traffic camera citations were defaulted on, and about 25% were eligible for debt collection. The randomised control trial demonstrated a 13% reduction in the likelihood of tickets defaulting and a 9% reduction in the likelihood of tickets ending up in debt collection, and the city implemented the new communications for the 600 000 tickets issued annually. At scale, initial results showed that, over the year, the new communications would lead to about 22 000 drivers – over a third of whom are people of colour – avoiding debt collection for an unpaid ticket.
Source: OECD/Bloomberg Philanthropies (2020), Survey on Innovation Capacity in Cities; What Works Cities (WWC) Certification.
Thus, data-driven assessment can increase public accountability by providing a tool for residents to hold governments responsible for budget and policy choices, and for governments to make adjustments based on tangible outcomes. But a significant gap between the interest in and measurement of local governments’ activity deprives the public of a data-driven accountability system. Such data could lead to policy learning, improvements to implementation of innovative programmes, and comparative analyses that identify the propensity for innovation across policy sectors, geography and city size.
Data governance framework for data-driven municipal governments
Data governance is a critical aspect of a data-driven government, and a good governance framework can help promote a common vision, formulate a coherent strategy and enhance the technical capacity to leverage data for residents’ well‑being. Data governance is not a technical or operational undertaking as much as a strategic and holistic framework considering every step of the government data value cycle, from collecting to processing, storing, publishing and using data. It also facilitates data sharing within and beyond the organisation, maximising the public value derived from the use and re-use of data.
Many organisations have data governance elements in response to technical and operational data challenges. However, fragmented elements can remain disjointed, contributing little to the strategic transformation of the organisation. Besides technical elements to ensure data interoperability and standards, an effective data governance framework also provides enabling conditions for the systematic and institutional use of data for problem solving and decision making. Most importantly, a data governance framework can lay a foundation for a public policy continuity that enables the organisation to move forward after changes in political leadership and administration.
At the national level, most OECD governments grapple to establish a framework that allows them to maximise the potential of data (OECD, 2019[9]). The OECD (2019[9]) notes that “data governance elements are often in place as part of broader digital transformation policies. However, these components can be fragmented, thus reducing their whole-of-government value in terms of public sector integration and cohesion.” For example, as part of the transformation into a data-driven and digital public sector, many OECD countries introduce regulations, standards and strategies for data management, digital government, open data or artificial intelligence. However, these tend to fall under fragmented governance arrangements, partly because different public sector organisations oversee different aspects of data. The fragmentation of internal organisation and governance hinders the integration and management of data. Fragmentation also stems from “legacy challenges in terms of what organisation generates and controls the data and the impossibility of sharing and accessing those data in light of specific legal arrangements.” (OECD, 2019[9]).
Recognising the range of national approaches, The Path to Becoming a Data-Driven Public Sector (OECD, 2019[9]) proposed a common framework for public sector data governance to clarify and standardise the concept and facilitate its effective implementation across countries. The framework builds on earlier versions in OECD digital government reviews such as OECD Digital Government Review of Norway (OECD, 2017[34]), OECD Digital Government Review of Sweden (OECD, 2019[35]), OECD Digital Government in Peru (OECD, 2019[36]) and the OECD Digital Government Review of Argentina (OECD, 2019[37]). Drawing on the OECD’s experience in digital government and government data, and extensive literature on data governance, the national framework organises (non-exclusive) data governance elements into six groups in three layers (Figure 3.5):
Strategic layer, including (A) Leadership and vision: The strategic layer regards data strategies and leadership roles as essential elements of good data governance. The framework purports that data strategies help ensure transparency and define leadership, expectations, roles and objectives.
Tactical layer, including (B) Capacity for coherent implementation, and (C) Regulation: The tactical layer draws on the public sector’s data skills, competencies, funding, collaboration and partnerships to generate public value from data. It also emphasises the importance of institutional networks, both formal and informal. The other group in this layer touches on the regulatory aspects of data, from technical and organisational standards to compliance with data-related rules and guidelines to ensure openness, protection, transparency and accountability.
Delivery layer, including the (D) Data value cycle, (E) Data infrastructure and (F) Data architecture: The delivery layer deals with daily implementation of organisational, sectoral, national and cross-border data strategies. This layer concerns technical and policy implications stemming from actions undertaken by different actors at various stages of the data value cycle. The delivery layer also touches on data infrastructure (i.e. adopting or adapting technological solutions such as APIs, cloud-based services, data lakes) and data architecture (e.g. standards, interoperability, semantics, etc.) to help public sector organisations achieve objectives defined in the strategic layer.
Figure 3.5. OECD framework for data governance in the public sector
Source: OECD (2019[9]) The Path to Becoming a Data-Driven Public Sector, https://dx.doi.org/10.1787/059814a7-en.
A well-conceived framework reconciles structure with agility. Besides providing a common language to gauge and benchmark progress, the OECD framework leaves ample room for “flexibility and scalability in order to avoid fragmentation; promote integration; and increase the adoption of good governance practices across organisations, levels of government, policy areas, sectors and border” (OECD, 2019[9]). While this framework initially meant to explore data practices from the national perspective, it holds potential and relevance for the sub-national level, where local governments seek to develop a comprehensive, coherent approach to data governance. These data governance elements are neither prescriptive nor exhaustive. Depending on organisational culture, local context and inherent challenges, municipalities might adopt or adapt them to realise their policy objectives. Applying a local lens to the previous framework, the framework tailored for data governance in the local public sector (Figure 3.6) aims to give municipalities a structured and holistic approach, as local governments often grapple with which areas to prioritise or how to embark on such journey. Certain domains and sub-domains of the tailored framework touch on data use practices in cities similar to those identified by the What Works Cities (WWC) Assessment programme. Therefore, we use survey results (as of April 2020) from the WWC Assessment programme to illustrate cities’ practices in leveraging data for city administration, policy design and evaluation.
Figure 3.6. A tailored framework for data governance in the local public sector
Source: Author’s adaptation from OECD (2019[9]), The Path to Becoming a Data-Driven Public Sector, OECD Digital Government Studies, https://doi.org/10.1787/059814a7-en.
Leadership and vision
Political leadership from mayors and elected officials is key to a culture that recognises data as a strategic asset to be enhanced, leveraged and shared, both across departments inside the administration and with external stakeholders. Data-driven governments value transparency and evidence-based decision making as their institutional priority. They are convinced of the potential value derived from data and data analytics, and seek to use these insights to guide their decisions. In recognizing that leadership is key, many data governance frameworks seek to reaffirm this prerequisite for local governments to maximise the use of their data. Convinced leadership can provide both the necessary financial commitment and the political support to overcome institutional resistance and silos between city agencies, enabling data initiatives to be maintained and even scaled up in a sustainable manner. More importantly, political leadership sets a vision and aspiration for ambitious data work, and long-term strategies to realise these goals.
Communicating and demonstrating the organisational focus of data-driven governance to municipal staff and external partners is essential, but the practice is not yet widespread in city leadership. The WWC Assessment shows that only 33% of mayors or chief executives communicate and demonstrate to staff that governing with evidence is an organisational expectation. Insufficient institutional backing can result in local public sector organisations adopting a risk-averse attitude towards data. The report written by the Auditor General for Wales, United Kingdom (2018[15]), found that many municipal governments “lack a vision, strategy or plan for improving data and are not clearly articulating what they need to do to improve.” Without a vision and strategy, data initiatives are left to individuals or disparate teams to navigate and push through the organisational resistance commonly found in local governments.
Because leadership empowers data-driven culture, a designated data leader and/or team can accelerate this transformation. At the national level, many OECD countries institutionalise the role of data stewardship within central/federal governments and across ministries. From a 2017 survey of OECD countries and partner economies, 11 out of the 34 mentioned that their central/federal government have a Chief Data Officer (CDO). At the sub-national level, similar functions are increasingly present among city governments and local public sector organisations over the past years (Figure 3.7). As of April 2020, among cities in the WWC Assessment programme, 58% have a data leader and 52% a team for developing and implementing performance management practices and data governance policies. When it comes to applying results-driven contracting strategies to upcoming procurements, contracts and/or grants citywide or within departments, cities also increasingly adopt the practice, with 25% of assessed cities putting in place a designated leader/team for the task.
Figure 3.7. Cities are increasingly designating leaders and/or teams for data policies and practices
Note: Out of 153 cities participating in the WWC Assessment programme as of April 2020, 88 cities meet the criterion for "Your local government has a designated leader and/or team responsible for developing and implementing citywide performance management practices" and 79 cities meet the criterion for "Your local government has a designated leader and/or team responsible for developing and implementing citywide data governance practices and policies". Meanwhile, only 39 cities meet the criterion for "Your local government has a designated leader and/or team responsible for applying results-driven contracting strategies to its portfolio of upcoming key procurements, contracts, and/or grants citywide or within departments".
Source: WWC Certification Database.
The creation of these roles to co-ordinate, synchronise and structure policy goals can build a more mature data-driven public sector (OECD, 2018[38]). Regardless of the levels of government, a designated leader and data team can spur interest in data use across organisations. Depending on the organisation, their responsibilities may vary from administratively managing data and related infrastructure to strategically enhancing the capacity of data analytics to improve well-being outcomes for residents; from establishing technical and organisational standards to ensuring compliance with data governance frameworks.
Besides knowledge creation, data leaders are often tasked with enhancing capacity for more systematic and extensive use of data. Though senior managers and municipal staff may be convinced by the public value derived from data use, many are not equipped with basic technical literacy and approach data use in a restrained manner. Some data leaders put in place city-wide data literacy training or help departments evaluate the effectiveness of their performance or interventions. One-third of cities participating in the WWC Assessment programme offer training to all local government staff on how to use data and evidence for decision making. For example, Enschede, Netherlands, invests in data awareness to facilitate internal data sharing, breaking down silos across departments (according to the OECD/Bloomberg Philanthropies Survey on Innovation Capacity 2020).
In short, a data leader ensures that an organisation approaches each stage of the government data value cycle in a strategic, efficient, user-friendly and compliant manner. In a report on the role of CDOs for the IBM Center for the Business of Government, Wiseman (2018[39]) enumerates various activities undertaken by CDOs and their peers who work on data (Table 3.2). While these are categorised based on their implications for either internal administration or external stakeholders, they tend to be “complementary and mutually reinforcing” (Wiseman, 2018[39]). For example, the quality of open data is largely determined by relevant data governance frameworks regulating the format, documentation and standards of published data sets.
Table 3.2. Governance, infrastructure and related activities of Chief Data Officers (CDO)
Note: 1. Tasks sometimes done by CDO and sometimes by other innovators.
Source: Wiseman (2018[39]), Data-Driven Government: The Role of Chief Data Officers, http://www.businessofgovernment.org.
Note that the designated data leader might go by titles other than CDO, such as Chief Data Scientist, Chief Data Analyst or Chief Information Officer. The role may also be embedded in or merged with other portfolios, leading to a variety of designations, including Head of ICT, Director of Innovation and Performance, Director of Statistics Office, Open Data Lead or IT Manager. While the responsibility of data leaders begins with data management and analytics, their designation may signal the organisation’s approach and priority areas for data use. For example, the Auditor General for Wales (2018[15]) found that many local governments in the country regard data leadership as either technical or legal in nature and assign the portfolio to the Head of ICT or Senior Information Risk Officer accordingly. Indeed, the role of data leadership is different from that of data administration. Data stewardship should not fall under the exclusive purview of IT or legal departments since it demands a strategic approach across dimensions including technical capacity, regulatory and legal frameworks, and organisational culture most importantly. Even though specialist experience and technical skills might be relevant, those positions might not serve as a fervent advocate seeking to transform the organisational culture and influence institutional attitudes towards a strategic use of data.
The Data Excellence Strategy for the City of Vienna (Austria), published in March 2019, represents a strong example of weaving strategic data use into the fabric of city administration. Vienna considers data the foundation for information and knowledge, essential to creating a “smart, intelligent and digital” metropolis. Its Data Excellence Strategy includes all necessary measures for the timely provision of reliable data that meets set quality standards, allowing the city to “provide reliable information and data as a central value of an open administration of the future,” thus creating benefits for residents, the economy, and science. Part of this commitment is the “Open by Default” guiding principle, requiring the city administration to open up publicly classified data, documents, and services in a machine-readable format and free of charge. (Digitales Wien, 2019[40])
Capacities for coherent implementation
Data skills and capabilities
Data skills and capabilities enable the public sector to develop insights for a variety of purposes. Data is extracted to measure, monitor, evaluate and assess a range of public policy, from programmes to services and experiments. From strategic planning to daily operations and targeted interventions, local governments rely on evidence derived from data to guide their decisions. Data are proactively gathered and integrated from sources across municipal agencies and non-governmental bodies to provide the city administration with a holistic picture.
Focusing on impact measurement, a mature, data-driven local government identifies how data can help evaluate the performance of city agencies and municipal services. Corrective changes based on insights derived from these data are taken to ensure a more efficient allocation of resources and continuous progress. Municipalities also put in place an organisational culture where experiments are conducted to optimise programmes that improve outcomes for residents. For every important decision, there should be actionable insights. In other words, municipal government should take advantage of data and analytics to anticipate future trends and risks, improve public service delivery and drive decision making (see “Unlocking the value of data for the local public sector”).
Nevertheless, the list above serves more as an aspiration rather than an expectation for many local governments. Besides organisational culture related to data use, the extent to which these activities are undertaken depends on factors such as the municipal staff’s commitment and internal capacity, and familiarity with and sophistication of data analysis, among others.
Box 3.7. The Lab@DC’s role in designing policy and programme interventions
The Lab@DC is a scientific team in the administration of the Mayor of the District of Columbia (DC), United States. The Lab@DC, in collaboration with various District agencies, leverages data and scientific methods to test, evaluate and improve public policies. To maximise the impact of their programmes, the Lab@DC uses a combination of approaches/methods such as randomised control trial, predictive modelling, resident-centred design and administrative data analysis.
From 2015–2017, the Lab@DC collaborated with the Metropolitan Police Department to explore whether police officers’ body-worn cameras (BWC) improved police-community interactions. Between June 2015 and December 2016, half of duty patrol and station officers were randomly assigned to wear BWCs, while the other half went without BWCs until December 2016. Outcomes were tracked until March 2017 using administrative data on documented police uses of forces, civilian complaints, policing activity and court outcomes. The study found that having a BWC had no substantial effect on police use of force, civilian complaints, policing activity or court outcomes. The results challenge conventional beliefs that BWCs can deter negative behaviours during police-community interactions. Such findings led to the implementation of new programmes aimed at improving police-community interactions such as a training that teaches the history and cultural context of DC. The Lab@DC is also evaluating the effectiveness of these new programmes.
Source: The Lab@DC (n.d.[41]), http://thelabprojects.dc.gov/.
City governments’ ability to govern with data relies on their capacity to draw insights from data. Despite recognising data as an asset, many cities still cannot adequately exploit it. Cities admit that they “need tools and expertise to close the gap between their intentions to use data in decision making and their actual capacity to do so” (What Works Cities, 2015[42]). According to a What Works Cities survey of US cities, 70% are committed to using data and evidence to make decisions about city programs, but only 28% modify existing programmes based on the results of data and evaluations (What Works Cities, 2015[42]). The main challenge for cities is to “build the requisite capacity and skills for collecting, storing, and analysing data in a depth and at a scale that are unprecedented, in addition to acquiring the infrastructure and computing power needed to store and process all the data.” (OECD, 2015[1]). Indeed, Figure 3.8 shows that limited data capacity hamper municipal efforts to optimise data, with 87% of cities lacking staff to support government innovation efforts and other activities (44% and 43% of these cities indicated the lack of staff to collect, store or analyse data as “Very challenging” and “Challenging” respectively). Edmonton, Canada, admits that high demand for skilled analysis – exceeding the capacity of in-house staff to collect, process, analyse and make sense of the massive datasets – has held back the city’s data use efforts.
Figure 3.8. Lack of data compatibility, staff capabilities and infrastructure challenge data use
Note: Out of 147 participating cities for both OECD/Bloomberg Philanthropies surveys, 138 responded to question “Which factors are the most challenging and prevent your municipality from optimising its use of data to support innovation goals?” Shown are some of the options that surveyed cities were asked to rank on a scale from 1 to 3 (1 = Very challenging, 2 = Challenging, 3 = Not a challenge, and N/A for “don’t know”).
Source: OECD/Bloomberg Philanthropies (2018-20), Survey on Innovation Capacity in Cities.
No single toolkit or strategy exists to enhance all municipal staffs’ data analytics capacity. Any activity would require one or a combination of types of data analysis, and thus varying data capabilities and skills. Some municipalities might be more interested in conducting randomised control trials to evaluate the effectiveness of targeted interventions and how to iterate them for better results. Others might be more concerned with results-driven contracting, conducting performance analysis for procurements, contracts and grants for either continuous or future improvement. Analysis can range from descriptive statistics to predictive models focused on forecasting and foresight, from simple stocktaking and data visualisation to advanced prescriptive analytics directly driving decisions. This needs local governments to be more proactive in identifying the skills gaps required for programmes and policy objectives.
While a dedicated data team and/or specialist staff might be entrusted with specific data analysis tasks, this does not negate the need for data literacy among non-specialist staff within the organisation. Best practices suggest that non-specialist employees should be able to understand how key decisions are made based on knowledge derived from data and communicate these findings to external partners and residents. Municipalities with more advanced data use capabilities can also provide partners (e.g. grantees, civic society, contracted vendors, etc.) with data literacy and skills training, conveying the expectation for data-driven and evidence-based governance. Once cities develop their capacity for data analysis, the insights from fine-grained data can enable municipal government to “better target infrastructure investments, deliver tailored public services and increase efficiency in operations and maintenance.” (OECD, 2015[1])
Even though the need to enhance data skills and capacity for municipal staff is universally acknowledged, it is not always addressed. Even though data analytics is valued as a core competence, many local governments consider the time and financial commitment for data training a significant obstacle given that professionals with data analytics skills are highly sought after by the private sector. This aspect of building data capacity can pose a particular challenge for small and mid-sized cities with tight budgets, diminished employment pools and limited access to partnerships. While the lack of data-related skills is common across sectors, it particularly affects stakeholders’ trust in the public sector’s capability and impedes their willingness to provide data.
Stakeholder engagement
Stakeholder engagement processes reveal the motives and activities of actors in both formal and informal institutional networks, which provides clarity, facilitates co-ordination and fosters trust. In this context, stakeholder engagement involves local governments’ efforts to include individuals and organisations concerned by data use in the processes of consultation, decision making and implementation. It should be noted that the community of stakeholders is diverse, having different and sometimes conflicting interests. Stakeholders can range from different levels of government, to residents (or service users), civil society and the private sector (e.g. enterprises, data-dependent start-ups, government-contracted companies, data providers, etc.). Understanding their characteristics would allow governments to adopt “more differentiated approaches to data access and sharing and a more effective management of the associated risks and incentives mechanisms.” (OECD, 2019[4]).
Depending on intentions, engagement with stakeholders can take different forms for various aspects of data use. The OECD (2015[43]) characterises the government-initiated stakeholder engagement process as a continuum of mechanisms that progresses from communication of information to “more intensive decision making where stakeholders exercise direct authority over the decisions taken”. The first level, communication, takes passive forms of information sharing and awareness raising. The sixth and most involved level of stakeholder engagement is co-production and co-decision, which entails a balanced share of power between actors over decision-making. Even though each level of stakeholder engagement can serve a different purpose and result in different impacts, collaborative and inclusive engagement helps governments identify priorities and address gaps in data policy and capability. Stakeholder engagement can inspire confidence that data systems and initiatives are working for the interests of society.
Figure 3.9. Levels of stakeholder engagement
Source: OECD (2015[43]), Stakeholder Engagement for Inclusive Water Governance, OECD Studies on Water, https://dx.doi.org/10.1787/9789264231122-en.
Besides communicating to external stakeholders about the benefits of these initiatives, many municipal governments try to build a community of data users and contributors to promote the values of sharing and re-using open government data. For example, for open government data, public organisations at all levels recognise that success relies on external stakeholders’ adoption of initiatives. Apart from making available a greater amount of open data online and maintaining user-friendly application programming interfaces (APIs), cities provide clear how-to guidance to help users access and utilise government data in a more effective manner. Many also gather information and seek feedback from open data users or track their applications to incorporate insights into the re-design and improvement of existing portals.
A vibrant community of open data users generates social and economic impacts from open government data and fosters trust in public institutions. As of April 2020, only 28 out of the 153 local governments in the WWC Assessment programme engage data users for the purpose of creating, revising, and/or improving the local government’s open data policies and practices. The number is higher for cities inviting community members to use public data to solve pressing community issues: 45 cities provide a process for partnership and collaboration with data users. However, only 23 cities (15%) meet both criteria for engaging open data users (see Figure 3.10).
Figure 3.10. Few cities have proactively and meaningfully engaged open data users
Note: Out of 153 cities participating in the WWC Assessment programme as of April 2020, only 23 cities meet both the criteria for "Your local government provides a clear process for partnership and collaboration with data users for the purpose of creating, revising, and/or improving the local government’s open data policies and practices" and "Your local government provides a clear process for partnership and collaboration with data users for the purpose of inviting community members to use public city data to solve pressing community issues".
Source: WWC Certification Database.
Stakeholder engagement processes can be resource-intensive in terms of both time and finance. Out of the approximately GBP 1 million that Transport for London, United Kingdom, spends annually on publishing open data, a sizeable portion goes to maintenance and engagement of communities of data users (Deloitte, 2017[44]; OECD, 2019[4]). This could explain why out of the 73% of cities from the WWC Assessment programme that publishes open data, only 35% go the extra mile to guide potential users in how to access and use these data.
Table 3.3. City governments that publish open data and/or provide how-to guidance to use city data
Note: Out of 153 cities participating in the WWC Assessment programme as of April 2020, 53 meet both criteria “Your local government publishes open data to a central, public online location.” and “Your local government provides clear how-to guidance to help residents access and use city data.”
Source: WWC Certification Database.
Engagement of data users can take an intensive and collaborative form of challenges or hackathons: competitive events where programmers, designers, data scientists, experts and interested individuals in various domains leverage government data to propose innovative technology solutions, an improve existing software and algorithms. Hackathons have proven to be a creative and useful method for governments to involve the community in addressing pressing issues while balancing risks by deciding the types of data and control mechanisms accessible to participants.
Stakeholder engagement can also take the form of partnerships to enrich existing data sets, co-produce databases and benefit from pooling resources and capabilities. Many local governments try to engage stakeholders upfront during the data production and sharing stage rather than simply expecting them to use data produced by the public sector. Data partnerships, be they public or public-private, can generate value that would be impossible to create when data is confined to a single organisation. Reimsbach-Kounatze (2015[45]) notes that the use and re-use of both public and private-sector data can enhance the power and quality of statistics, especially in a global context where national surveys for data collection are losing momentum. Statistics offices and data teams can tap into various sources of non-official data held by private companies to produce more robust indicators, enrich their insights and improve their evidence-based decision making in fast-changing areas such as urban planning, crisis management, etc.
As Figure 3.11 shows, most cities take advantage of data partnerships (specifically aimed at enhancing innovation capacity). While the number vary across US and European cities, only 13% of cities surveyed report that data partnerships play an insignificant role. More than three-quarters of participating cities indicate that they collaborate with academia, think tanks and research institutions to collect and analyse data. However, with fewer than half reporting some form of partnership with either the private sector or private philanthropy, cities may want to take more advantage of such opportunities.
Figure 3.11. Cities establish a wide range of partnership to enhance their data work
Note: Out of 147 participating cities for both OECD/Bloomberg Philanthropies surveys, 142 responded to the question "Has your municipality developed any partnerships with the aim of collecting or analysing data to fuel innovation capacity or strategy?" Surveyed cities were asked to select all options that apply. The vertical axis denotes the percentage of municipalities developing external partnerships for their data work.
Source: OECD/Bloomberg Philanthropies (2018-20), Survey on Innovation Capacity in Cities.
Public-private partnership (PPP) for data refers to long-term, voluntary agreements between the public sector (e.g. government and public agencies at various levels) and private partners to enhance the capacity to derive benefits from such data for the parties involved. Different from ad-hoc and one-off data-sharing initiatives, PPPs are characterised by “the existence of an agreement which structures, collaboration and defined roles, responsibilities and rights.” (Robin, Klein and Jütting, 2016[46]). Long-term collaboration is important when longitudinal non-official data from the private sector are often required to complement traditional sources of statistics collected by the public sector. While PPPs for data and statistics have existed for quite some time, the emergence of Big Data and the private sector’s ability to process it has renewed attention to public-private data partnerships.
PPPs in data and statistics can provide governments with new and granular data in a timely and cost-effective manner. Another advantage of PPPs is that a municipality cab tap into the competences, skills and technologies of partners to perform advanced analysis from holistic sources of data. Through these collaborations, city government can establish a network of stakeholders whose support and expertise can be leveraged to deepen the impact of data. Like other forms of engagement, data partnerships can be cost-intensive as joint agreements must maximise the potential of co-operation while juggling the conflicting interests of different partners. Any agreements related to PPPs for data must be pre-defined with the involved parties’ responsibilities and potential liability clearly structured. In cases where institutions exchange data (especially potentially sensitive data), strong regulatory mechanisms must be in place to prevent and resolve misuse. Also, PPPs for data are a two-way process where private companies need to be convinced of the benefits of entering into these agreements.
Box 3.8. Select examples of data partnerships and initiatives developed by cities
Bologna, Italy, is working with Barcelona, Spain, to learn from the experience of their Data Office. The cities are working to define a collaborative model of governance for data generated and managed by public and private stakeholders. Through the Fondazione Innovazione Urbana created by the municipality and the University of Bologna, the city established a Data Office involved in collection and management of data generated by the city’s participatory processes.
Bristol, United Kingdom, collaborated with seven local authorities (Gloucestershire, South Gloucestershire, Wiltshire, Somerset, North Somerset, Bath and North East Somerset, and Devon) to develop a digital heritage mapping resource called “Know Your Place”. With access to a range of historic maps and data, “Know Your Place” lets users explore and add information about their local areas and neighbourhoods, contributing to a rich and diverse community map of local heritage for everyone.
Edmonton, Canada, is part of the Metrolab network, which cultivates partnerships between universities and local governments to drive evidence-based policy and enable data and technology transformation. The city also formed partnerships with academics (especially through student projects) and with non-profit agencies (e.g. social service sectors).
Helsinki, Finland, has partnerships with regional water and waste management agencies, public transport agencies and hospitals (that control a large amount of healthcare data).
Rosario, Argentina, in partnership with the provincial government of Santa Fe and academic institutions, established the technological pole (Polo Tecnologico) of Rosario to nurture a culture of innovation and promote technological development locally and internationally. The city collaborates with universities (e.g. University of Gran Rosario) and non-profit civil associations (e.g. International Association of Educating Cities).
Tempe, AZ, United States, in partnership with Arizona State University’s Biodesign Institute, uses wastewater analytics to collect data on opioid abuse in their community and inform where to send resources on education and overdose response. The city is now working to use these practices to determine COVID-19 cases.
Source: OECD/Bloomberg Philanthropies (2020), Survey on Innovation Capacity in Cities; What Works Cities (2020[47]), Data-Rich Sewage in Tempe, AZ, https://medium.com/what-works-cities-certification/data-rich-sewage-in-tempe-az-77b2444a23f8; The City of Tempe (2021[48]), “Innovation in Advancing Community Health and Fighting COVID-19”, https://covid19.tempe.gov.
Data openness
The What Works Cities Assessment Glossary defines open data as “electronic data records that are accessible in whole or in part to the public and are legally open without restriction on use or re-use. This practice is a form of proactive disclosure – making information available without it being requested.” Open data can come from various actors in the data ecosystem such as individuals, households, private companies and public sector organisations. The degree of “openness” (which is a spectrum more than a binary) depends on how the data are generated, and which actors possess and decide to open them to wider application. Most definitions of open data establish the types of data that should be made available and lay out characteristics for “openness”. For instance, data openness can be deliberated based on criteria such as accessibility, machine readability (i.e. compatible formats that make data easily retrieved and processed), costs (i.e. free of charge) and rights (i.e. free of restrictions on intellectual property rights to use and distribute) (McKinsey&Company, 2014[49]).
In the absence of conflicting interests, it is widely expected that public sector data be accessible as open government data. Central to such an expectation is the conviction that public sector information in government data systems, once personally de-identified, should be open to benefit society. Open data is the most prominent and widely adopted approach by governments and public entities when it comes to enhanced access and data sharing (OECD, 2019[4]). This is because the public sector is one of the most data-intensive sectors. For example, US public sector agencies stored 1.3 petabytes (PB) of data on average in 2011, making them the country’s fifth most data-intensive sector (OECD, 2015[1]). By 2020, the US federal government’s open data portal had published over 200 000 data sets, together with guides and other resources for users to conduct research, develop applications and design data visualisations.
In an era of declining public trust, local governments worldwide have considered the open data movement as an opportunity to foster transparency and public participation. Residents’ use of open data can help “increase openness, transparency and accountability of government activities and thus boost public trust in governments” (OECD, 2019[4]).
An open data initiative that help bring to light unaccountable government activities and irresponsible public spending is the Brazilian Transparency Portal, created in 2004 to increase the fiscal transparency of the federal government. Built on a collaboration between ministries and federal organisations, the open government budget data portal discloses previously secret information on federal agencies’ and government officials’ expenditures, and companies blacklisted from government contracts. Since its establishment, the Portal significantly contributed to the country’s anti-corruption efforts, garnering more than 900 000 unique visitors every month (Graft, Verhulst and Young, 2016[50]). More importantly, the model of open government budget data inspired transparency initiatives throughout local governments in Brazil and other Latin American countries. However, similar trends were not widely observed at the local level, especially when it comes to public contracting transparency and accountability.
As of April 2020, only 13 out of the 153 local governments in the WWC Assessment programme proactively share information about contracts, procurement, and/or vendor performance to increase bid competitiveness and ensure transparency and accountability (see Figure 3.12). Indeed, the problem with government data remains that they tend to be closely guarded, and access is granted on an ad-hoc basis. Data-sharing, be it externally or internally, faces institutional and bureaucratic resistance, leaving it undervalued, underdeveloped and underutilised.
Figure 3.12. Few cities make data available to increase bid competitiveness and strengthen procurement transparency
Note: Out of 153 cities participating in the WWC Assessment programme as of April 2020, only 13 meet the criterion for "Your local government proactively shares data, documents, and information about contracts, procurement, and/or vendor performance, in order to increase bid competitiveness and strengthen procurement transparency and accountability".
Source: WWC Certification Database.
Open government data is a step toward transparency and accountability, but the quality of data and its potential to generate public value matter at least as much as the quantity (Janssen et al., 2017[24]; Kuk and Davies, 2011[51]). Excess data is no substitute for high-quality or user-friendly data. For example, budgetary information, while being released as open data, can remain unintelligible to most concerned residents since it requires a certain level of technical understanding. Despite having a well-functioning open data portal, Bloomington, IN, United States, acknowledged that the lack of data curation and visualisation limits wider application of the service, both internally and externally (from the OECD/Bloomberg Philanthropies Survey on Innovation Capacity 2020). Therefore, the use of open data to increase government transparency can only be effective if these efforts are accompanied by “additional measures for enhancing government accountability and transparency, as well as democratic control” (OECD, 2015[1]).
Local governments should be aware that the implications of open data as “a strong focus on transparency, though essential to sustain efforts meant to strengthen overall public sector integrity and accountability, can limit the proactive release of open government data and the necessary engagement of the relevant actors in the ecosystem in data reuse for value creation.” (OECD, 2018[38]). Data released out of concern for accountability might prompt governments to open data in a more passive or reactive manner. A government can open its data and remain deeply non-transparent and unaccountable. Non-discriminatory and purposeful release of local government data can help maximise the social and economic impact of these data, generating public value beyond public trust and government transparency.
Open data also presents business opportunities for other actors and stakeholders in the data ecosystem, such as start-ups and companies relying on data to develop innovative commercial and social goods and services. In a sense, this can be considered “indirect intervention” supporting economic and entrepreneurial activities “recycle” government data to produce public or social goods.
At the local level, the tangible socio-economic benefits of open data were confirmed in a 2017 (Deloitte[44]) report on Transport for London (TfL), United Kingdom. TfL’s open data produced a virtuous cycle for London’s transport network providers and users (e.g. GBP 130 million in economic benefits for London, TfL and its customers, and road users) as well as positive externalities such as improved air quality job creation, and a boon to the innovation ecosystem (Table 3.4). The open data provided by TfL facilitates cross-sector co‑operation, allowing for better integration between transport and navigation services (e.g. integrating disparate transport modes and route planning options). Such social and economic impacts suggest that the use (and re-use) of public sector data, most prominently in the form of open government data, is a major enabling condition for open innovation.
Table 3.4. The social and economic benefits of TfL’s open data
Based on data provided by TfL, the Deloitte report estimated the economic benefits and cost-savings for three core segments: passengers (all network users), London and TfL itself.
Source: Deloitte (2017[44]), Assessing the value of TfL’s open data and digital partnerships, http://content.tfl.gov.uk/deloitte-report-tfl-open-data.pdf.
At the municipal level, open data initiatives can only live up to their potential when there is demand both within and outside the administration. The Auditor General for Wales (2018[15]) report confirms that the lack of data skills is an obstacle to maximising open data. Governments must go beyond publishing open data and develop capacity to perform the main functions of the data cycle – collecting data, opening and sharing data, combining data (i.e. ensuring compatibility), and analysing data for new insights and applications – resulting in innovation and action-oriented decision making (Janssen et al., 2017[24]). This includes the need to establish and invest in data literacy for all city staff, not just specialised teams. As discussed in “The professional background of innovation teams may reflect cities’ priorities for innovation” section of Chapter 2, developing data competency across a public administration can facilitate data-driven decision-making in all aspects, making it second nature among staff and reducing skill asymmetries that create cross-departmental friction. Despite the difficulties cities face in releasing their data to increase transparency and promote data-driven innovation, the investment in open data is worthwhile.
Box 3.9. Leveraging data use for public sector innovation
Data can serve as an important input for innovation activity. Public sector organisations that seek to improve their innovative capacity should focus on three main data-related aspects:
Sourcing: The identification of different types and sources of data, information and knowledge that are relevant. This may also involve explicit efforts to generate new knowledge.
Exploiting: Organisations need to channel data, information and knowledge into a usable form so that it can be fully exploited to support evidence-based decision making and organisational renewal (to support the development of “learning organisations”).
Sharing: Organisations need to share information collected with wider sets of actors including other public-sector organisations and members of the public to support decision making, accountability and co-innovation and facilitate value creation elsewhere in the economy.
Source: OECD (2015[52]), The Innovation Imperative: Contributing to Productivity, Growth and Well-Being, https://dx.doi.org/10.1787/9789264239814-en.
Legal and regulatory frameworks
An important element of the data governance model, legal and regulatory frameworks range from national legislative measures to softer instruments such as guidelines and recommendations issued by various levels of government. Legal and regulatory frameworks help cities “define, drive and ensure compliance with the rules and policies guiding data management, including data openness, protection and sharing.” (OECD, 2019[9]) At the same time, however, these measures might create a barrier to good data governance since fragmented, inflexible and incoherent legal and regulatory frameworks can hamper data-sharing efforts, delay data integration and impede the management of data.
It is imperative that legal and regulatory frameworks facilitate data exchange, both horizontally among city agencies and vertically among levels of government. This is because organisational siloes, which legally preclude data sharing and data compatibility, can hinder the potential of data use for timely decision making. Guidelines and instructions for horizontal and vertical data sharing can help municipalities challenge such silos in public administration and encourage co-operation among jurisdictions and other levels of government, which has “long been recognised as a crucial element of efficient and effective urban governance” (OECD, 2015[1]; Rodrigo, Allio and Andres-Amo, 2009[53]).
An effective regulatory framework that allows for a flexible data sharing agreement would embed mechanisms to strengthen co-ordination across different municipal organisations and facilitate public-private partnerships on urban data. This would provide local governments with an opportunity to revamp and streamline internal procedures and delivery services in a transformative manner. For example, a proactive data-sharing agreement would prevent duplicate data being requested multiple times, relieving the administrative burden on both municipal employees and service users. Secondly, cross-sectoral data sharing by public sector organisations would allow data to be linked, integrated and leveraged for performance analysis and decision making.
When it comes to open government data, its success depends on the local government’s ability to put in place stringent regulations to shield publicly available data from privacy and cybersecurity threats. Many local governments have guidelines and frameworks to address the security, privacy and ethical dimensions of open data. However, while a legal framework for data protection can safeguard the rights of data subjects and limit security breaches, it might inadvertently limit the flows of data and hinder local governments’ willingness to open their data.
Patterns in city data use and residents’ well-being outcomes
Data use is a tool for local governments to safeguard and enhance the well-being of their residents. In this sense, city practices regarding data use should be examined with respect to their potential to influence residents’ well-being outcomes. As is the case for local public sector innovation capacity (see Chapter 2), the relationship between data use by city governments and residents’ well-being can be complex. Some well-being outcomes, like education, facilitate the use of data, while others, like housing affordability and city satisfaction, can, in turn, be favourably affected by data practices.
This section explores the links between data use practices and well-being outcomes at the city level. It builds on the OECD well-being framework for regions and cities (Figure 1.1) and on information about data use in cities participating in the What Works Cities (WWC) Assessment. As explained in Chapter 1, the measurement of data use practices in cities relies on the WWC Standard’s 45 criteria of excellence for data use in local government, grouped into eight foundational areas: Data Governance, Evaluations, General Management, Open Data, Performance and Analytics, Repurposing, Results-Driven Contracting, and Stakeholder Engagement (see Box 1.2). Information on the WWC’s 45 criteria at the city level is used to create a score from 0 to 45 that captures cities’ efforts in using data for city administration, policy design and evaluation.
The section first looks at how the cities that underwent the WWC Assessment are doing across 11 well-being dimensions based on their level of data use practices. It then explores links at the city level (controlling for population and economic development) between well-being indicators and the different foundational areas of data use.
The results suggest robust correlations between different data use areas – notably stakeholder engagement, open data and performance and analytics – and the well-being indicators of city and life satisfaction, educational attainment, material conditions, affordability of housing and self‑reported health. While these associations do not imply causality from data use to well-being, they suggest that data use capacity tends to accompany improvements in several dimensions of people’s lives.
How is life in cities with high data use capacity?
Cities with more advanced data use practices tend to show better well-being outcomes. Looking at 145 cities (141 US cities and 4 EU cities), high data use standards are associated with better outcomes in 10 out of 11 well-being dimensions, on average. What is more, large differences (above 4 percentage points) exist between cities with high data use standards (23–38 practices) and low (0–9 practices) in seven out of ten well-being dimensions, such as health, civic engagement, income, access to services, education, city satisfaction and life satisfaction (Figure 3.13 and Figure 3.14). The only exception to this pattern is in the dimension of safety – measured in terms of crime rates – where cities with low data use practices display better results, on average. One possible explanation for this (although not demonstrated in this report due to data constraints) is that cities with high crime rates invest more in monitoring and data‑generating infrastructure (such as cameras).
It is also worth highlighting that these results do not consider population and income effects, which tend to be associated with, both, data use practices (Annex Table 3.A.1) and well-being outcomes. The following section highlights the links between data use and well-being outcomes that remain robust even when considering city size and income.
Figure 3.13. Data use and well-being in cities: Indexes
Difference between indexes ranging from 0 to 100, where 100 is the best outcome; unweighted averages of cities
Note: High data use cities: from 23 to 38 practices; Medium data use cities: from 10 to 22 practices; Low data use cities: from 0 to 9 practices. Indicators included in the indexes by well-being dimension are listed in Table 1.1. Number of cities in parentheses.
Sources: Data used to compute the indexes comes from the American Community Survey; Gallup US Daily 2016-17; NYU Langone Health and NYU Wagner City Health Dashboard; Reflective Democracy Campaign; and WWC Certification database, 2018-20.
Figure 3.14. Data use and well-being in cities: Headline indicators
Note: The table presents the average of each indicator for Low data use, Medium data use, and High data use cities, prior to the transformation into the normalised indexes presented in Figure 3.13. The units of each indicator appear in the first column in parentheses.
Sources: American Community Survey; Gallup US Daily 2016-17; NYU Langone Health and NYU Wagner City Health Dashboard; Reflective Democracy Campaign; and WWC Certification database, 2018-20.
Data use practices in cities and residents’ well-being outcomes
People in cities with higher data use standards are more likely to be satisfied with their city and their life, on average. While 81% of people living in cities with high data use capacity are satisfied with their lives, only 77.5% of people in cities with low data use report being similarly satisfied with their life (Figure 3.15). The positive correlation between higher data use standards and life satisfaction persists even after controlling for city size and economic conditions (see 0), two of the factors that are likely to affect life satisfaction, among other well-being outcomes.
Figure 3.15. Data use and satisfaction with the city and life
Note: The positive association between having medium or high data use standards and life and city satisfaction persists after controlling for population and the level of economic development of the city.
Sources: Gallup US Daily 2016-17; European Survey of Quality of Life, 2019; and WWC Certification database, 2018-20.
Well-implemented data practices by city governments can positively affect several residents’ outcomes related to city and life satisfaction, including health, affordability of housing and overall material conditions. Because of the interlinkages across well-being dimensions, data practices targeting a particular well-being aspect can spill over into other dimensions. For example, an open data platform that leads to the development of new apps helping residents access preventive health services can improve health outcomes, which in turn can affect life satisfaction. The findings, presented in the next paragraphs, based on a large sample of cities in the United States support the argument that data use potentially enables city and life satisfaction.
Data use practices that enhance service delivery, increase transparency and accountability of the local government, or that promote co-creation with stakeholders seem to drive many of the links observed to well-being outcomes. These practices are mainly contained in three out of the eight foundational areas defined by the WWC Standard: Stakeholder Engagement, Open Data, and Performance and Analytics.
Engaging residents to design data policies is key to better well-being outcomes. Resident engagement and feedback can help ensure continuity of innovative public projects (Arundel and Es-Sadki, 2019[54]). Similarly, stakeholder engagement activities (like civic hackathons) can help put data into public use and provide local governments with valuable information affecting future data releases and policies (Robinson and Johnson, 2016[55]). More generally, residents who have a say in the way their local governments act and use data can experience higher well-being through both subjective feelings (e.g. their satisfaction with their city, government and life) and objective outcomes (e.g. better housing opportunities) (OECD, 2015[56]).
In the cities studied, stakeholder engagement practices showed robust and positive correlations with better life satisfaction and housing outcomes (Figure 3.16). Within the WWC framework, the foundational area of stakeholder engagement captures local governments’ commitment to engage with data users in the design and implementation of data policies and practices. Objective well-being outcomes, like housing, may be influenced by stakeholder engagement in data use practices if they lead to better-designed policies or to better access to government programmes (e.g. student or mono-parental aid, access to social housing). In turn, subjective well-being outcomes, like life satisfaction, may be influenced by stakeholder engagement practices either directly, such as if exchanges with the city administration led to an increased sense of community, or indirectly, such as by improvements in interlinked objective well-being outcomes.
Figure 3.16. Stakeholder engagement vs. Life satisfaction and Housing affordability
Note: The housing affordability indicator looks at households without mortgages. The positive association between stakeholder engagement practices and the life satisfaction and housing affordability indicators persists after controlling for population and the level of economic development of the city.
Sources: American Community Survey; and WWC Certification database, 2018-20.
Similarly, practices that encourage and facilitate the publication of open data can improve residents’ well-being. Beyond the direct value of information, open data is likely to produce positive externalities across well-being dimensions. Open data can be used to build tools that improve access to services, monitor the activities of local governments and create data products that inform residents. One of the clearest links is to city satisfaction. Local governments that engage in open data practices tend to have a higher share of residents satisfied with the city (Figure 3.17, Panel A), suggesting that residents value the increase in transparency and accountability of their local governments. Housing affordability – measured by the share of people spending less than 25% of their income on rent – also shows a positive correlation with open data practices (Figure 3.17, Panel B). Although the open data measures refer to access to data at large (not by sector), this effect could still be driven by better access to information on the housing market (e.g. supply and prices) and housing aid, including social housing. In addition, there are examples where city data on housing, particularly home ownership, leads to more targeted funding for the construction of family rental units, which subsequently improves housing affordability (What Works Cities, 2019[57]).
Figure 3.17. Open data vs. City satisfaction and Housing affordability
Note: The housing affordability indicator looks at households without mortgages. The positive association between open data practices and the city satisfaction and housing affordability indicators persists after controlling for population and the level of economic development of the city.
Sources: Gallup US Daily 2016-17; European Survey of Quality of Life, 2019; and WWC Certification database, 2018-20.
City governments that use data to evaluate and monitor progress toward specific goals are more likely to provide residents with better services, resulting in improved well-being outcomes such as self-reported health. The Integrated Medical Information and Analytical System (IMIAS) project in Moscow, Russia, is an example of how local governments can monitor the performance of their programmes to bring about better health outcomes (Box 3.5). The project provides authorities real-time metrics on the number of patients, waiting times, length of visits and estimated cost savings, helping satisfy residents’ demand for medical services (Moscow Mayor official website, 2016[58]). This is consistent with the outcomes observed in the WWC city sample: city governments with more data use practices in Performance and Analytics had, on average, a higher share of residents reporting no health problems (Figure 3.18, Panel A). Although modest, this correlation is robust even after correcting for the effect of city size and residents’ income.
Beyond the health dimension, high data use capacity in Performance and Analytics can allow residents to enjoy better material conditions through better public service delivery, even for people with the same income (Figure 3.18, Panel B). Overall, based on the analysis of sample cities, 78% of people living in cities with high standards in data use (above 22 practices) report having enough money to get things they need, compared to 73% in cities with low standards in data use (less than 10 practices). Improvement in the material conditions of residents can happen when there is monitoring and evaluation of programmes that give residents access to better public transport, affordable housing and social aid. For example, Barcelona, Spain, monitors and directs assistance to populations at risk of social exclusion and who do not have access to public aid programmes offered by the city (Box 3.6). Similarly, the London Borough of Barking and Dagenham, United Kingdom, uses predictive analytics to identify and help households at risk of homelessness (Box 3.4).
Figure 3.18. Performance and Analytics vs. Self-reported health and Material conditions
Note: The positive association between performance and analytics practices and the self-reported indicators on health and material conditions persists after controlling for population and the level of economic development of the city.
Sources: Gallup US Daily 2016-17; and WWC Certification database, 2018-20.
Combining data use and PSI capacity in cities
PSI and data use practices might go hand in hand and reinforce each other, affecting well‑being outcomes such as city and life satisfaction. While previous sections examine and document the correlations between well-being outcomes and either PSI capacity (Chapter 2) or data use (Chapter 3) separately, this section provides evidence on the associations between data use and PSI practices, and on the potential reinforcing effects that combining them might have on city and life satisfaction, based on a sample of 57 cities that participated in both the OECD/Bloomberg Philanthropies Survey on Innovation Capacity in Cities and in the WWC Assessment.
Cities with a formal innovation strategy, experienced innovation staff and funding for innovation tend to exhibit higher capacity in data use at the local level. Overall, cities with good PSI practices tend to have higher standards in data use (based on the WWC score), and the correlation coefficient between the PSI and data use scores is 0.3, statistically significant at 95%. In addition, the differences in data use scores are particularly high – above 8 percentage points – between cities with and without a formal innovation strategy and a staff with at least five years of experience (Figure 3.19).
Figure 3.19. Average data use capacity by PSI practice
Sources: OECD/Bloomberg Philanthropies Survey on Innovation Capacity in Cities, 2018-20; and WWC Certification database, 2018-20.
Cities with high PSI and high data use capacity have higher city and life satisfaction than cities with lower levels of PSI and data use. Beyond the reinforcing effects that PSI and data use might have on each other, cities with high PSI and data use also display better well-being outcomes than cities with low PSI and data use. For example, in the surveyed cities, around 86% of people in cities with both high PSI and high data use report being satisfied or very satisfied with their city, almost 8 percentage points above cities with low PSI and low data use. The same pattern holds for life satisfaction, where the gap between the best and worst performers in terms of PSI and data use is around 4 percentage points (Figure 3.20).
Figure 3.20. The intersection of data use and PSI vs. city and life satisfaction
Sources: Gallup US Daily 2016-17; OECD/Bloomberg Philanthropies Survey on Innovation Capacity in Cities, 2018-20; and WWC Certification database, 2018-20.
References
[54] Arundel, A. and N. Es-Sadki (2019), “D2.7 Preliminary survey results report: Understanding value co-creation in public services for transforming European public administrations”, https://www.co-val.eu/public-deliverables/ (accessed on 4 February 2021).
[15] Auditor General for Wales (2018), The maturity of local government in use of data, Auditor General for Wales, https://www.audit.wales/sites/default/files/Maturity-of-local-government-in-use-of-data-ENG_6.pdf (accessed on 2 December 2020).
[20] Basu, M. (2016), Jakarta’s plans for predictive government, GovInsider, https://govinsider.asia/innovation/jakartas-plans-for-predictive-government/ (accessed on 13 April 2021).
[25] City of New Orleans (n.d.), Code Enforcement Abatement Tool: A NOLAlytics Project.
[29] Cohen, B., E. Almirall and H. Chesbrough (2016), “The City as a Lab”, California Management Review, Vol. 59/1, pp. 5-13, http://dx.doi.org/10.1177/0008125616683951.
[44] Deloitte (2017), Assessing the value of TfL’s open data and digital partnerships, http://content.tfl.gov.uk/deloitte-report-tfl-open-data.pdf.
[27] Deloitte (2013), Market Assessment of Public Sector Information, Department for Business, Innovation and Skills, London, https://assets.publishing.service.gov.uk/government/uploads/system/uploads/attachment_data/file/198905/bis-13-743-market-assessment-of-public-sector-information.pdf.
[40] Digitales Wien (2019), Data Excellence Strategy of the City of Vienna - Digitales Wien, https://digitales.wien.gv.at/en/data-excellence-strategy-of-the-city-of-vienna/ (accessed on 5 August 2021).
[14] European Commission (2019), Monheim Am Rhein steps up the challenge to deliver fully digital trusted public services, European Commission, https://ec.europa.eu/cefdigital/wiki/display/CEFDIGITAL/2019/08/05/Monheim+Am+Rhein+steps+up+the+challenge+to+deliver+fully+digital+trusted+public+services.
[12] Freed, B. (2019), One year after Atlanta’s ransomware attack, the city says it’s transforming its technology, https://statescoop.com/one-year-after-atlantas-ransomware-attack-the-city-says-its-transforming-its-technology/.
[33] Gault, F. (2018), “Defining and measuring innovation in all sectors of the economy”, Research Policy, Vol. 47/3, pp. 617-622, http://dx.doi.org/10.1016/j.respol.2018.01.007.
[49] Gold, A. and E. Hannigan (eds.) (2014), Innovation in local government: Open Data and Information Technology, https://www.mckinsey.com/~/media/mckinsey/dotcom/client_service/public%20sector/gdnt/gdnt_innovation%20in%20local%20gov_open%20data_it_full_book.ashx.
[50] Graft, A., S. Verhulst and A. Young (2016), Brazil’s Open Budget Transparency Portal: Making Public How Public Money Is Spent, https://odimpact.org/case-brazils-open-budget-transparency-portal.html#footnote-020.
[21] Hillenbrand, K. (2016), Predicting Fire Risk: From New Orleans to a Nationwide Tool, https://datasmart.ash.harvard.edu/news/article/predicting-fire-risk-from-new-orleans-to-a-nationwide-tool-846.
[8] ISO (2017), “ISO/IEC 19941:2017 - Information technology — Cloud computing — Interoperability and portability”, International Organization for Standardization, https://www.iso.org/standard/66639.html (accessed on 13 April 2021).
[24] Janssen, M. et al. (2017), Driving public sector innovation using big and open linked data (BOLD), Springer New York LLC, http://dx.doi.org/10.1007/s10796-017-9746-2.
[51] Kuk, G. and T. Davies (2011), “The roles of agency and artifacts in assembling open data complementarities”, International Conference on Information Systems 2011, ICIS 2011. 2.
[11] Midena, M. (2020), “Les 15 Propositions De France Digitale Pour “Réarmer Technologiquement” Le Pays”, Forbes France, https://www.forbes.fr/technologie/les-15-propositions-de-france-digitale-pour-rearmer-technologiquement-le-pays/ (accessed on 13 April 2021).
[58] Moscow Mayor official website (2016), Integrated Medical Information and Analytical System: Taking Moscow healthcare online, https://www.mos.ru/en/news/item/13864073/ (accessed on 3 February 2021).
[7] Narayanan, A. and V. Shmatikov (2006), “How To Break Anonymity of the Netflix Prize Dataset”, Cornell University, http://arxiv.org/abs/cs/0610105.
[31] OECD (2020), Innovation for Development Impact: Lessons from the OECD Development Assistance Committee, The Development Dimension, OECD Publishing, Paris, https://dx.doi.org/10.1787/a9be77b3-en.
[36] OECD (2019), Digital Government in Peru: Working Closely with Citizens, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/0c1eb85b-en.
[37] OECD (2019), Digital Government Review of Argentina: Accelerating the Digitalisation of the Public Sector, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/354732cc-en.
[35] OECD (2019), Digital Government Review of Sweden: Towards a Data-driven Public Sector, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/4daf932b-en.
[4] OECD (2019), Enhancing Access to and Sharing of Data: Reconciling Risks and Benefits for Data Re-use across Societies, OECD Publishing, Paris, https://dx.doi.org/10.1787/276aaca8-en.
[9] OECD (2019), The Path to Becoming a Data-Driven Public Sector, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/059814a7-en.
[38] OECD (2018), Open Government Data Report: Enhancing Policy Maturity for Sustainable Impact, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264305847-en.
[34] OECD (2017), Digital Government Review of Norway: Boosting the Digital Transformation of the Public Sector, OECD Digital Government Studies, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264279742-en.
[1] OECD (2015), Data-Driven Innovation: Big Data for Growth and Well-Being, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264229358-en.
[56] OECD (2015), How’s Life? 2015: Measuring Well-being, OECD Publishing, Paris, https://dx.doi.org/10.1787/how_life-2015-en.
[43] OECD (2015), Stakeholder Engagement for Inclusive Water Governance, OECD Studies on Water, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264231122-en.
[52] OECD (2015), The Innovation Imperative: Contributing to Productivity, Growth and Well-Being, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264239814-en.
[5] OECD (2013), Recommendation of the Council OECD Legal Instruments concerning Guidelines Governing the Protection of Privacy and Transborder Flows of Personal Data, OECD, Paris, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0188.
[22] OECD (2012), OECD Internet Economy Outlook 2012, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264086463-en.
[23] OECD (2010), Cities and Climate Change, OECD Publishing, Paris, https://dx.doi.org/10.1787/9789264091375-en.
[2] OECD (2008), Recommendation of the Council for Enhanced Access and More Effective Use of Public Sector Information, OECD, Paris, https://legalinstruments.oecd.org/en/instruments/OECD-LEGAL-0362#dates.
[32] OECD/Eurostat (2018), Oslo Manual 2018: Guidelines for Collecting, Reporting and Using Data on Innovation, 4th Edition, The Measurement of Scientific, Technological and Innovation Activities, OECD Publishing, Paris/Eurostat, Luxembourg, https://dx.doi.org/10.1787/9789264304604-en.
[6] Ohm, P. (2009), The Rise and Fall of Invasive ISP Surveillance, https://papers.ssrn.com/abstract=1261344.
[45] Reimsbach-Kounatze, C. (2015), “The Proliferation of ’Big Data’ and Implications for Official Statistics and Statistical Agencies: A Preliminary Analysis”, OECD Digital Economy Papers, No. 245, OECD Publishing, Paris, https://dx.doi.org/10.1787/5js7t9wqzvg8-en.
[26] Results for America (2018), How to clean city streets? Los Angeles begins by collecting new data, https://results4america.org/wp-content/uploads/2018/12/LosAngelesCaseStudy_Final.pdf (accessed on 22 April 2021).
[46] Robin, N., T. Klein and J. Jütting (2016), “Public-Private Partnerships for Statistics: Lessons Learned, Future Steps: A focus on the use of non-official data sources for national statistics and public policy”, OECD Development Co-operation Working Papers, No. 27, OECD Publishing, Paris, https://dx.doi.org/10.1787/5jm3nqp1g8wf-en.
[55] Robinson, P. and P. Johnson (2016), “Civic Hackathons: New Terrain for Local Government-Citizen Interaction?”, Urban Planning, Vol. 1/2, pp. 65-74, http://dx.doi.org/10.17645/up.v1i2.627.
[53] Rodrigo, D., L. Allio and P. Andres-Amo (2009), “Multi-Level Regulatory Governance: Policies, Institutions and Tools for Regulatory Quality and Policy Coherence”, OECD Working Papers on Public Governance, No. 13, OECD Publishing, Paris, https://dx.doi.org/10.1787/224074617147.
[16] Salge, T. and A. Vera (2012), “Benefiting from Public Sector Innovation: The Moderating Role of Customer and Learning Orientation”, Public Administration Review, Vol. 72/4, pp. 550-559, http://dx.doi.org/10.1111/j.1540-6210.2012.02529.x.
[13] Sneed, A. (2019), What Cities Can Learn From Atlanta’s Cyberattack, Bloomberg CityLab, https://www.bloomberg.com/news/articles/2019-10-29/what-cities-can-learn-from-atlanta-s-cyberattack.
[19] The Behavioural Insights Team (2017), The Behavioural Insights Team Update Report 2016-17, Behavioural Insights Team, https://www.bi.team/publications/the-behavioural-insights-team-update-report-2016-17/.
[48] The City of Tempe (2021), Innovation in Advancing Community Health and Fighting COVID-19, https://covid19.tempe.gov/ (accessed on 22 April 2021).
[41] The Lab@DC (n.d.), The Lab@DC, http://thelabprojects.dc.gov/ (accessed on 13 April 2021).
[17] Toderas, A. and M. Manning (2019), The Future of Predictive Analytics in Councils.
[18] UK Local Government Association (2020), Using predictive analytics in local public services, https://www.local.gov.uk/using-predictive-analytics-local-public-services.
[10] van Ooijen, C., B. Ubaldi and B. Welby (2019), “A data-driven public sector: Enabling the strategic use of data for productive, inclusive and trustworthy governance”, OECD Working Papers on Public Governance, No. 33, OECD Publishing, Paris, https://dx.doi.org/10.1787/09ab162c-en.
[3] Verhulst, S., A. Young and P. Srinivasan (n.d.), An Introduction to Data Collaboratives: Creating Public Value by Exchanging Data, https://datacollaboratives.org/introduction.html#section1.
[28] Vickery, G. (2011), Review of recent studies on PSI reuse and related market developments, European Commission, https://ec.europa.eu/digital-single-market/en/news/review-recent-studies-psi-reuse-and-related-market-developments.
[30] Vries, H., V. Bekkers and L. Tummers (2016), “Innovation in the public sector: A systematic review and future research agenda”, Public Administration, Vol. 94/1, pp. 146-166, http://dx.doi.org/10.1111/padm.12209.
[47] What Works Cities (2020), Data-Rich Sewage in Tempe, AZ. 2020, https://medium.com/what-works-cities-certification/data-rich-sewage-in-tempe-az-77b2444a23f8 (accessed on 22 April 2021).
[57] What Works Cities (2019), Data Is Just the Beginning: It’s about Better Serving Residents, https://whatworkscities.medium.com/data-is-just-the-beginning-its-about-better-serving-residents-98eb0e9cc56.
[42] What Works Cities (2015), What Works Cities Brief: The City Hall Data Gap, https://data.bloomberglp.com/dotorg/sites/8/2016/03/WWC_Brief_c.pdf.
[39] Wiseman, J. (2018), Data-Driven Government: The Role of Chief Data Officers, http://www.businessofgovernment.org.
Annex 3.A. Data use practices by city size and income
Annex Table 3.A.1. Adoption of data use practices by city size and income
Note: Sample of 145 cities. Number of data practices in parentheses.
Sources: WWC Certification database, 2018-20.
Annex 3.B. Regression results: Well-being outcomes and data use practices
Annex Table 3.B.1. Regression results: Selected well-being outcomes and data use practices
Notes: Linear regressions using ordinary least squares (OLS) estimation. Standard errors are corrected for heteroscedasticity. t statistics in parentheses. * p < 0.10, ** p < 0.05, *** p < 0.01.
Sources: American Community Survey; Gallup US Daily 2016-17; and WWC Certification database, 2018-20.