Governments are drawing on traditional and non-traditional data sources to guide public service design, implementation and improvements. They are increasingly using experimental tools to help them navigate highly complex and unpredictable environments.

Following a series of flash floods in 2011, Singapore began to create a 3D map of the city-state. The world’s most densely populated nation, Singapore wanted to identify the areas most at risk of flooding and make better use of its land. Launched in 2014 and completed in 2022, Virtual Singapore is a 3D digital model of Singapore that uses real-time and topographical data. It is a digital twin of the city-state and the first digital twin of a country. This tool also facilitates collaboration among government agencies, utility companies and other stakeholders, enabling informed decision-making and resource allocation. Additionally, real-time data integration enhances emergency services’ response planning and improves transportation efficiency. This innovation offers immense value in enhancing urban planning, infrastructure management and disaster preparedness. It allows decision-makers to optimise land use, assess flood risk and manage underground utilities more effectively, harnessing the power of granular evidence to support better public services.

Governments are moving beyond the use of traditional data sources, such as internal records, official statistics and surveys, which have certain drawbacks, such as infrequent updates, narrow scopes, self-reporting bias or inconsistent coverage (Bertoni et al., 2023[40]). Service performance data has been lacking or under-utilised and was not used in decision making. Rather, data was often used afterward as a control factor (Maciejewski, 2016[41]). The ability of traditional data sources to inform decision-making was limited and therefore limited public services’ agility: their ability to improve, quickly learn from feedback, integrate knowledge and change to better respond to people’s needs.

Governments increasingly explore and use new data sources, including experimentation and simulations, to improve their service design, management, delivery and evaluation. These data insights allow better management of services throughout their lifecycle. Developing administrative and technical structure to obtain actionable and usable data can enable real-time and distributed and iterative service evaluation (Höchtl, Parycek and Schöllhammer, 2015[42]). Notable examples of novel approaches that support decision-making to improve public services are Türkiye’s Public Services Monitoring System (see Box 5.1) and Togo’s Embedded Evidence Lab, which aims to optimise the use of collected data to assist public decision-making and to make social protection programmes more effective.

Non-traditional and unstructured data sources, such as sensor, satellite imagery and big textual data, enable governments to obtain richer, real-time insights that complement existing monitoring mechanisms and reduce time lags associated with conventional feedback loops. Among the relevant innovations in this area are the United Kingdom’s InnOvaTe Programme (see Box 5.2), which uses some 3,000 sensors and a centralised data hub to monitor different public spaces and services; Korea’s approach to IoT (see Case Study 4), focusing on mobility and pedestrians’ safety; and, Sierra Leone’s Integrated Geographical System Portal (IGIS) cross-sectoral spatial data infrastructure, which helps the government implement policies and interventions towards the Sustainable Development Goals. In Singapore, whole-of-government platforms for the analysis of unstructured textual data, which use natural language processing tools, speech-to-text transcription and video analytics, make it possible for every public servant to analyse and make sense of non-traditional data.

Governments are using open data initiatives, which foster transparency, accountability and citizen engagement, to improve public services. By sharing data on service performance and policy-relevant activities, governments allow citizens, researchers and businesses to access valuable information that can drive innovation and improve service outcomes. Open data allows for a collaborative approach to problem-solving, where external stakeholders can contribute insights and solutions based on real-time, accessible data. Open data platforms can help track the efficiency of public transportation systems, monitor environmental quality or evaluate the impact of social programmes. Examples such as Mexico City’s Open Data Portal, Czechia’s Data KHK and France’s PILOTE demonstrate how releasing government data can enhance information accessibility and transparency, and lead to more data-driven decision-making processes that improve public services.

Governments are increasingly using experimentation to improve public service delivery and decision-making. This allows public institutions to create new design approaches, test hypotheses and prototypes, and assess the impact of different interventions in controlled environments (such as pilots, testbeds and innovation labs) before implementing them on a larger scale. For instance, the United Kingdom's What Works Network has been instrumental in evaluating social programmes through randomized controlled trials (RCTs) to inform policy decisions in areas such as education and crime reduction. In the United States, Los Angeles initiated  BIG:LEAP, a Universal Basic Income pilot intended to break the cycle of intergenerational poverty, and measured the programme’s effects through RCTs. Portugal's Public services as living lab initiative has helped to transform real-life public services into safe spaces that engage public servants, citizens and researchers and can be used to explore, prototype and pilot in successive loops potential new solutions to improve public services. Additionally, Canada’s Experimentation Works initiative has showcased experiments that allow public servants to develop and execute projects with support from experts, fostering a culture of evidence-based governance through iterative learning and open sharing of results.

Governments are also adopting simulations, such as digital twins, which allow them to explore risk-taking or even disruptive initiatives to service design and delivery, while controlling costs and creating a learning environment, as Singapore has done. Digital twins can simulate a variety of relevant social activities and can thus inform public services. For example, they can be developed to avoid traffic congestion or simulate emergent social segregation, and they can represent the complexity of these systems (Caldarelli et al., 2023[43]). For example, the digital twin of Patras, Greece, which provides analysis, citizen feedback data and advanced technologies, represents a comprehensive approach to urban management and decision-making (Gkontzis et al., 2024[44]). In Spain, Barcelona has adopted a digital twin to assess the proximity of services and facilities, and help plan and design metro lines.

The OECD Recommendation on Human-Centred Public Administrative Services puts forward that by making dedicated efforts to collect and analyse data, governments can obtain a holistic and inclusive view of public services that is crucial to their continuous improvement (OECD, 2024[27]). However, implementing these innovative data-driven approaches in public services can face several challenges. First, governments may not have the know-how and skills necessary to effectively harness and interpret complex data sources, such as sensors and satellite imagery. Second, changes in political leadership risk disrupting the continuity and sustainability of these initiatives (especially when they are set up as pilots). Finally, transparency and open data initiatives can complicate government efforts to ensure data privacy and security. By fostering a culture of continuous learning and adaptation, leveraging experimental approaches to evidence-based decision-making, governments can overcome these obstacles.

Good practices for harnessing the power of data:

• Governments should tie data and innovations to the decision-making process in systematic, usable and actionable ways. Public organisations can approach the ongoing monitoring and evaluation of services with a view of ensuring continuous improvements based on evidence. For instance, the use of non-traditional data sources and techniques, such as open and publicly available digital footprints (e.g. social media), or the real-time display of structured information to provide decision makers usable data offers promising paths for public organisations to explore.

• Engaging with people is crucial and the datafication of public spaces and decision-making should not overlook users, who are the origin of data and are impacted most by the decisions. Users may not accept innovations that do not take them into account, which could compromise their trust in public organisations.

• Robust policies and infrastructure are needed to balance transparency with the protection of sensitive information.

• Governments should make targeted and dedicated investments to measure results and impact of innovations. It is important that new evidence on services includes information on service quality and user satisfaction, and that it is leveraged to understand cost-benefits of innovations, the return on investment for government and value for money.

In the Seongdong District of Seoul, South Korea, public transportation services had deteriorated due to extreme climate events, including severe temperatures and fine dust, which made public transport unsafe and unattractive. Previous attempts to address these issues (such as sun shelters and wind-blocking bus stops) often lacked a coordinated strategy and resources, which hindered their impact.

To respond to these issues, Seongdong District turned to the Internet of Things (IoT) – a network of interconnected devices that collect and exchange data through internet – and non-traditional data sources to enhance transportation services and pedestrian safety. It introduced Smart Shelter, futuristic bus stops that run on smart technologies and collect data for 24-hour real-time monitoring of traffic information, temperature and air quality. Smart Shelters are also equipped with smart services and safety systems, such as intelligent CCTV and an emergency bell. The features and information collected allow public authorities to respond immediately in the event of an emergency, for example.

To ensure that Smart Shelter met people’s needs, Seongdong District asked residents to participate in a Living Lab – a participatory platform that collects citizens’ ideas for improving services – and to share their ideas for upgrading the existing bus stops. During co-design sessions, residents identified service gaps, and suggested ideas and designs for transportation services according to their needs.

Seongdong District has installed 52 Smart Shelters and plans to add more, which would expand the service beyond 2.09 million users annually. The Smart Shelters have led to a decrease in traffic accidents and a decline in crime rates, demonstrating how the innovation can make residents safer. Furthermore, residents gave Smart Shelters a 95% satisfaction rating.

Smart Shelters have received international media coverage and inspired more than 20 local governments to adopt similar initiatives. Adopting new features such as solar panels, an uninterruptible power supply and revenue from advertisements, the innovation is gradually becoming self-sustainable. Efforts are also being made to make the shelters more inclusive, automated external defibrillators (AEDs) will be installed for improved safety, along with hearing loops to assist the hearing-impaired.