BIG DATA in ACTION for GOVERNMENT Big Data Innovation in Public Services, Policy and Engagement Contents Introduction .............................................................................. 1 Strengthening Government Service Delivery...................... 3 Smarter Policymaking............................................................. 7 Citizen Engagement ................................................................ 11 Putting Big Data into Action for Government .................... 14 Additional Resources.............................................................. 16 Introduction Big data is a term widely used to performance and innovation in service delivery and policymaking processes. The UK’s Digital describe the exponential growth of Government Unit estimates the country’s data, particularly the data flowing government can save around US$20-41 billion annually through big data analytics – a figure from ubiquitous mobile phones, based on efficiency gains, reduced fraud satellites, ground sensors, vehicles and error, and improved tax collection alone. McKinsey estimates big data can reduce the and social media. It also explains the European Union’s administrative costs by rise of the computing technologies and 15 to 20 percent, creating the equivalent of at least US$223-446 billion in new value1. In algorithms that harness big data for developing countries, governments have an valuable insights. In the public sector, opportunity to adopt big data solutions and leapfrog traditional administrative approaches. big data typically refers to the use of non-traditional data sources and Big data is a pressing issue, particularly data innovations to make government at a time when many are concerned about the role of information in political change. solutions more responsive and effective. Governments in the United Kingdom, the United States, Singapore and Korea are among Governments have an opportunity to harness many2 that have adopted high-level big data big data solutions to improve productivity, strategies. These frameworks focus on the 1. See http://www.mckinsey.com/business-functions/digital-mckinsey/ our-insights/public-sector-digitization-the-trillion-dollar-challenge || http:// www.mckinsey.com/business-functions/digital-mckinsey/our-insights/big- data-the-next-frontier-for-innovation 2. Other countries with big data strategies include Australia, France, Japan, New Zealand BIG DATA in ACTION for GOVERNMENT 1 various roles of government as producer, US Census Bureau buys commercial data consumer and facilitator of big data to for address verification, for example, and enable better service and policy outcomes: transportation agencies use information from GPS navigation systems to inform • Producer: More than a million datasets from traffic management. governments around the world are available online, from tax returns and unemployment • Facilitator: Governments should claims to hospital funding and energy use. invest in big data research and stewardship, as well as establish • Consumer: Governments can use big data robust legal and technical frameworks (both their own and from other sources) to ensure big data is accessible and to promote responsive government. The responsibly used for public good. P R OMI SI NG APPL I CATION S F OR BIG DATA IN GOVE RN ME N T The potential for big data to transform government is vast. This brief focuses on big data solutions with applications in service delivery, policymaking and citizen engagement – areas where big data can play a transformational role. It also discusses key initiatives where the government needs to facilitate action for effective data use. SERVICE DELIVERY POLICY MAKING CITIZEN ENGAGEMENT Big data analytics Policymakers are By applying can be used by using satellite machine learning governments to imagery, cell phone to online and improve existing data and more to social media, services and to produce alternative governments can be draw on novel economic indicators more responsive to datasets to drive for new – and citizen sentiment, entirely new real-time – policy ushering in a new public services insights dimension of civic engagement Application Areas for BIG DATA in Government 2 BIG DATA in ACTION for GOVRNMENT Strengthening Government Service Delivery These days, consumers are PROCUREMENT AND FINANCIAL FLOWS: Big data techniques are transforming processes accustomed to commercial to detect fraud, collusion and money laundering products and services growing through the analysis of procurement and financial information. Algorithms are used more personalized and always to trawl procurement records, invoices, bank on-demand. People interact with information, certificates of origin and other data government services every day in to identify fraud, collusion and shell companies. health, employment, education and business, just to name a few. Society Solution Spotlight: The International Consortium of Investigative Journalism (ICIJ) provides nearly is eager to use big data to make 40 years of information on companies in over 200 public service delivery equally as countries and territories. ICIJ hosts the Mossack smart, responsive and personalized. Fonseca (aka Panama) papers – 11.5 million emails, database entries, images and documents which big data analytics helped to rapidly sort, categorize, One area of tremendous opportunity is index and search. The analysis found several in the use big data algorithms to detect irregularities and triggered multiple investigations. fraud, non-compliance and bottlenecks in government operations. Governments are huge producers of data, most of which Solution Spotlight: The Counter Fraud Centre is unstructured and text-intensive. Text at the UK’s Chartered Institute of Public analytics and machine learning algorithms Finance and Accountancy (CIPFA) manages are indispensable to analyze administrative the London Counter Fraud Hub, which supplies data for insights. They can automate big data analytics, investigations and recovery systematic, multilevel checks on insurance, services for the 32 London Local Authorities procurement and tax records to flag entries and the City of London Corporation. requiring further examination, or to identify performance bottlenecks that need attention. BIG DATA in ACTION for GOVERNMENT 3 HEALTHCARE: Valuable information about Solution Spotlight: During West Africa’s 2014 public health is found in government records, Ebola crisis, Swedish nonprofit Flowminder as well as unconventional sources such as used cellphone data to create models social media, discussion forums, mailing predicting population mobility and potential lists, health websites and news outlets. routes of disease transmission, while Mobile phone data is also being widely experts from around the world created the used for monitoring and improving health Ebola Open Data Initiative. This organized services. These channels can complement access to multiple big data resources and traditional sources to help authorities created a tool to analyze them, leading to the monitor results and feedback in real-time, development of an open-source global model and improve performance and outcomes in for Ebola, which allowed health officials to hospital management operations, insurance test different scenarios and interventions. processing, vaccinations, nutrition and disease. PUBLIC UTILITIES: Remote sensing data Solution Spotlight: HealthMap is a platform from satellites and ground-based sensors that automatically tracks and analyzes can provide a wealth of real-time or near-real- multiple data-feeds in 15 languages – time information to monitor the provision including official documents, news and quality of public utilities such as water reports, social media, international health and energy. organizations, government websites and health workers’ blogs – to produce an online visualization of disease trends. Solution Spotlight: Shanghai’s municipal This approach can facilitate the early government has installed sensors detection of disease outbreaks, enabling throughout the city’s 3,300-kilometer prompt and targeted responses. waterpipe network. These sensors allow the authorities to identify precise locations of issues such Solution Spotlight: in Brazil, the Ceara State as leaks, dramatically reducing government is using Proactive listening, response times, infrastructure a mobile-phone based system to provide damage, maintenance costs the government with real-time data on and service disruption. the coverage and quality of public health services. Real-time information enables the government to identify emerging problems, such as bribes for health services, and intervene in a timely and effective manner. 4 BIG DATA in ACTION for GOVRNMENT Solution Spotlight: The India.Nightlights EDUCATION: Personal data from devices, platform uses night light data from satellite exam data and other sources can be used images to monitor electricity provision in innovative ways to monitor student over time to all 600,000 villages in India. performance, better understand teaching practices, and help parents and students identify the best fit with a school. Big data Solution Spotlight: The SWEET Lab is working can complement traditional sources to enable with local governments and service providers new delivery regimes that tailor lessons in developing countries to understand to performance, and monitor progress water and sanitation services. Sensors towards government and global goals. on water pumps in Kenya and Ethiopia provide usage statistics and monitor functionality, to help keep pumps working. Solution Spotlight: In Kenya, the national government uses a mapping platform to show areas where educational resources are lacking. Tanzania has established the website Shule.info to help assess the quality of schools. Solution Spotlight: In Mexico, the ENOVA platform that uses data analytics and data from student interaction and feedback to continuously troubleshoot educational processes and improve learning tools. BIG DATA in ACTION for GOVERNMENT 5 LAND ADMINISTRATION: High-resolution PUBLIC SECURITY: Police forces are drawing imagery from satellites and Unmanned on big data and predictive analytics to make Aerial Vehicles (UAVs or drones), combined better policing decisions. Basic information with cloud computing and handheld such as crime type and location can help computing devices, can deliver capabilities officers make smarter decisions on patrol. to revolutionize land registration processes Law enforcement agencies are also using big and survey methods. These near-real-time data analytics, through social media analytics methods will inevitably replace traditional companies such as Crimson Hexagon, to be surveying approaches. Land registration more effective in public outreach campaigns. improves public revenues by increasing tax collection, enables efficient resolution of competing boundary claims and provides Solution Spotlight: PredPol is a predictive the unbanked with property rights that are policing application used in 50 cities the basis for access to credit and finance. that analyzes basic data, such as crime type, dates and location, to help law enforcement agencies make better Solution Spotlight: In Kosovo, UAVs are being policing decisions. Brazil, the UK and the tested to secure property rights. The drone Netherlands use similar approaches. imagery is processed in real-time and used in combination with a handheld tablet to record boundaries in an online system to formally Solution Spotlight: In Bogota, Colombia, register land. The entire process is completed World Bank researchers are using big in hours and days, as opposed to weeks and data analytics and risk terrain modeling to months with traditional survey methods. understand relationships between crime and public infrastructure such as bus stations, public hospitals, schools and drugstores. Solution Spotlight: The South African government has worked with Digital Globe to use high-frequency and high-resolution satellite imagery feeds along with mobile phone applications to geo-locate each residential dwelling with an address. This is part of a broad government effort to digitalize census and household survey processes for planning and budgeting. 6 BIG DATA in ACTION for GOVRNMENT Smarter Policymaking The role of information in policymaking big data can potentially deliver policy insights on a more frequent, disaggregated and cost- is a big concern in our information age. effective basis. Traditional surveys – like Big data is a viable source of high- censuses, official statistics and enterprise data – will always be needed, but it is likely frequency and granular data that can that robust big data proxies will continue to provide profound insights into human emerge and become more integrated into policymaking processes and decisions. The mobility and economic behavior, to data from satellites, mobile phones and better inform policy decisions. The entire social media alone can change the dynamic between information and policymaking. globe is now imaged by satellites on a daily basis, with quality and affordability TRANSPORT AND URBAN PLANNING: Satellite imagery, cell phones, vehicle sensors, video continuously improving. New public- feeds and social media are being used by private arrangements, like the Open policymakers and planners for traffic and urban Transport Partnership and the Workforce planning. Remote sensing, mobile phone data and machine learning can provide policymakers Data Initiative, are emerging to make and planners with much better understanding data from social media, professional of urban mobility, land usage and urban change. Policymakers now and in the future networks, mobile phones and sensors can use these insights to provide access readily available for policymakers. to jobs and make cities more sustainable, inclusive, productive, resilient and livable. Traditionally, policy is based on reports that are largely informed by traditional survey and Solution Spotlight: Big Pixel is among administrative data and statistics – instruments several satellite based applications that are slow and labor intensive. In contrast, that are enabling policy makers to have BIG DATA in ACTION for GOVERNMENT 7 systematic understanding of where OFFICIAL STATISTICS: In many countries, urbanization is occurring, to what degree census data is incomplete and outdated. and how quickly, showing what impact Accurate and reliable information on basic policy interventions like zone changes and statistics such as population and poverty transport corridors have upon urbanization. are surprisingly lacking for much of the world. A consensus is emerging in the statistics community that big data and Solution Spotlight: OpenTraffic is a global machine learning can be used in certain platform that uses anonymized positional instances to create viable statistical proxies data from vehicles and smartphones to for monitoring and policy decisions. visualize real-time and historic traffic conditions. The Uber Movement platform Solution Spotlight: Facebook Connectivity Lab aggregates Uber vehicle data to inform is working in collaboration with the Center urban infrastructure decisions. for International Earth Science Information Network to produce high-resolution population Maps. Flowminder is using Solution Spotlight: In Seoul, South Korea, satellite, mobile phone and household survey the OWL bus uses big data analytics data to create dynamic population maps. to better serve the needs of nighttime travelers. The project used data from three billion call and text data points, along Solution Spotlight: The Billion Prices with five billion points from corporate and Project (BPP) started at the MIT Sloan private taxi data, to design nighttime bus School of Management aggregates routes optimized to match the origin and price information from a multitude of destination of passengers’ journeys. online retailers around the world and gives real-time inflation predictions. Solution Spotlight: Stanford’s sustainability and artificial intelligence lab has developed a satellite imagery and deep learning approach to predict poverty. The World Bank is also experimenting with satellite- based methods that complement traditional methods for small-area poverty mapping. 8 BIG DATA in ACTION for GOVRNMENT FOOD SECURITY: Big data tools are enabling EMPLOYMENT SERVICES: Government sophisticated analysis of weather and labor agencies are experimenting with satellite data to analyze crop strategies and big data to inform the most appropriate formulate optimal agricultural policies. policies to help individuals back to work, such as tailoring training services for different segments of job seekers. Solution Spotlight: Lobell Labs has developed a scalable crop yield mapper that uses satellite-based measurements Solution Spotlight: LinkedIn is working with to predict crop yields in both commercial the Australian government to mine data and smallholder farms. The solution from the LinkedIn economic graph to identify has been tested in the US, as well as in trends such as the increasing demand for smallholder systems in Africa and India. technology workers to also possess soft skills. These insights will help policymakers to create programs giving Australian Solution Spotlight: Information on food workers the right skills and competencies. prices can be crowdsourced through social media. In Indonesia, Global Pulse has tested social media and crowdsourcing to track Solution Spotlight: The World Bank is food commodity prices in near-real-time in working with Flowminder in Haiti to areas where other data sources are limited. analyze mobile phone data to provide Haitians with better information on and access to jobs and opportunities. BIG DATA in ACTION for GOVERNMENT 9 ENVIRONMENTAL POLICY: Analysis of Solution Spotlight: Microsoft’s Madingley is environmental and climate data from multiple a next-generation model of ecoystems and sources is enabling authorities to understand biodiversity across the globe. Madingley environmental impacts and interventions, from provides a working simulation of the regional, national and global perspectives. global carbon cycle, and aims to model everything from deforestation to animal migration, pollution and overfishing, Solution Spotlight: Global Forest Watch in a real-time “virtual biosphere.” combines satellite imagery, crowdsourced witness accounts and public datasets to track deforestation around the world, believed to be a leading manmade cause of climate change. The project helps ethical businesses to ensure that their supply chain is not complicit in deforestation. 10 BIG DATA in ACTION for GOVRNMENT Citizen Engagement Citizen engagement – the interaction and improve the transaction experience. Benefits include: between government and citizens to strengthen citizen voice – can play • A more informed and engaged electorate: Informing voters can help them assess an integral role in the improvement politicians’ performance and legislative of service delivery and policymaking issues of interest, and can increase voter turnout. processes, as long as appropriate • Monitoring and feedback on services and mechanisms are in place to translate polices: Big data can be a catalyst to citizen feedback into action. Big data improve service delivery and proactively listen to feedback on performance and analytics can make this interaction quality, as well as to gauge sentiment smarter, more targeted, personalized around policies and interventions. • Citizen voice and collective action: Big data and responsive. analytics help to identify real-time trends and mobilize pressure on policymakers. Governments have only recently begun to adopt the customer relationship management approaches that are now prevalent in the The promise of big data for better citizen private sector in many countries. These engagement can only be realized if feedback is systems can potentially be used to enhance effectively aligned with government incentives, government-citizen interactions with timely mechanisms and processes to take informed information and to enable citizens to better action. The pursuit of public ends is not just manage service requests. Big data analytics a responsibility of governments, but also of can also improve government process the private sector, nonprofits and the public. performance, enable better decisions Individual administrations within the different BIG DATA in ACTION for GOVERNMENT 11 systems of government will embrace the idea of shared greater accountability to varying degrees. Making government services and policies more responsive to citizen feedback will require strong relationships between policymakers, service providers, civic organizations and citizens. Solution Spotlight: After Haiti’s 2010 earthquake, a crowdsourced application developed by Humanitarian Open Street Map soon became the default tool for search and rescue teams. More than 600 volunteers traced roads and encampments from aerial images into a computer program. Crowdsourced markers identified resources such as refugee camps and cholera clinics. Solution Spotlight: In Brazil, researchers analyzed tweets made during protests surrounding the 2014 Soccer World Cup, to understand citizens’ feelings about policies and government. The analysis showed citizens expressing negative sentiment about the national government’s low investment in basic services compared to lavish spending on the World Cup – suggesting domestic policy priorities for Brazil’s government. Solution Spotlight: The Ushahidi platform was created to monitor Kenya’s general elections. It has since been used around the globe by citizens to monitor and support solutions for elections, civic problems and crises. 12 BIG DATA in ACTION for GOVRNMENT BIG DATA in ACTION for GOVERNMENT 13 Putting BIG DATA into Action for Government The age of big data creates new perseverance as solutions are tested, adapted and mainstreamed – but the potential value challenges and opportunities. from effectively using big data should make the Governments need to develop strategies, effort worthwhile. tools and forms of engagement to better Big data has the potential to help shape understand dynamic forces and respond government of the future. However, leveraging the opportunities brought about by big data accordingly. The solutions featured also requires a cultural shift within the public in this brief show how big data can sector. Data needs to be seen as an asset rather than a byproduct of administrative tackle fraud and corruption, generate activity in order to be valued, curated and administrative savings, and improve shared where appropriate. This could be part of a broader movement to transform the public service delivery and policymaking sector away from the bureaucratic model processes, making them smarter, more inherited from the 19th century and towards accountable and more responsive to a more agile, 21st century version. Such a transformation could be key to restoring citizen citizen feedback. trust in government, which has significantly eroded in recent years. However, to realize value from big data, governments must strengthen technical and legal Several key areas require government action frameworks to access and use data responsibly. if big data is to be used effectively for better Importantly, they need to develop integrated government: capabilities to put big data insights into action, and to be responsive to citizen feedback on 1. Provide clear regulation and guidelines for services and polices. Big data innovation is data use: Governments need to provide a rich process of learning that will require legal and policy guidelines on data 14 BIG DATA in ACTION for GOVRNMENT ownership, quality and sharing, privacy, 3. Promote transparency in algorithms: civil liberties and equality. They must make Governments must encourage good decisions on privacy in conjunction with practices in the design and documentation public opinion, informing citizens about the of algorithms, including transparency and trade-offs between the privacy and security accountability mechanisms to correct and risks of sharing data, versus the benefits. contest decisions based on algorithms. 2. Foster Public-Private Partnerships: The 4. Invest in big data capacities: Many majority of high-value data assets reside datasets span departments, meaning with the private sector. Public-private governments need to manage and partnerships can coordinate governments, integrate datasets from departmental businesses and nonprofits to make big silos. Big-data ‘control centers’ are data accessible for the public good. needed for implementing strategies and The Global Partnership for Sustainable best practices, and providing essential Development Data is a multi-stakeholder resources in advanced computing network of over 150 organizations to technologies. Governments also need promote the sharing and responsible use change-management and data science of private-sector data. skills to adapt process, systems and practices to put big data in action. BIG DATA in ACTION for GOVERNMENT 15 Additional Resources Useful starting points on use of big data to improve government include: Big Data Challenge, Pioneering Data Driven The Federal Big Data Research and Development Approaches to Development: World Bank report Strategic Plan: Written by the Big Data Senior that profiles the application and lessons of several Steering Group, Subcommittee on Networking and World Bank big data applications in the development Information Technology Research and Development, sector. for the National Science and Technology Council, Executive Office of the President, United Sates. United Nations Global Pulse: is an initiative to promote awareness of opportunities to “harness big Pathways for Progress, Data Driven Development: data for development and humanitarian action.” It’s a World Economic Forum report on vision of how Big network of innovation labs that coordinate research Data might be used to address the challenges of and “strengthen the big data innovation ecosystem.” sustainable development. UN International Conference on Big Data for Official Bringing probability judgments into policy debates Statistics 2016: Proceedings of the 2016 conference via forecasting tournaments: A tournament in Dublin. “Going beyond the existing examples of sponsored by the U.S. intelligence community isolated use of mobile phone data, social media revealed ways in which forecasters can better data or satellite imagery data, innovations are use probability estimates to make predictions— needed in the daily production of official statistics, even for seemingly “unique” events—and showed which requires real partnerships with the private that tournaments are a useful tool for generating sector, new skills and infrastructure, and clear links knowledge. between these available Big Data sources and the SDG indicators. Beyond prediction: Using big data for policy problems: Machine-learning prediction methods UN Big Data Global Working Group: Big Data have been extremely productive in applications Project Inventory: a catalog of Big Data projects ranging from medicine to allocating fire and health that are relevant for official statistics needed for inspectors in cities. However, there are a number decision-making on public policies, as well as of gaps between making a prediction and making for management and monitoring of public sector a decision, and underlying assumptions need to programs/projects.” It is a joint project of the World be understood in order to optimize data-driven Bank and United Nations Statistics Division. decision-making. Big Data Applications in the Government Sector: World Development Report, 2016: Digital Dividends: Communications of the ACM [Association for report look at digital opportunities of the new Computing Machinery], March 2014. Good overview economy and how countries can prepare by of business and government use of big data, ensuring that institutions are accountable. including issues of volume, velocity, variety, and security. The public sector is learning from fast- World Development Report, 2017, Governance and growing commercial use of big data and analytics. the Law: report looks at fundamental questions, which are at the heart of development, policy making Big Data: A Report on Algorithmic Systems, and policy implementation. Opportunity, and Civil Rights: Executive Office of the President, May 2016. “This report examines several case studies from the spheres of credit and lending, hiring and employment, higher education, and criminal justice to provide snapshots of opportunities and dangers, as well as ways that government policies can work to harness the power of big data and avoid discriminatory outcomes.” 16 BIG DATA in ACTION for GOVRNMENT Contact This solutions brief is an input into the 2017 World Government Summit workshop and high-level panel on Big Data for Government. It was produced in collaboration with the World Bank Governance Practice and Innovations in Big Data Analytics program, which resides within the Global Operations and Knowledge Management unit of the World Bank. For additional information, please contact Trevor Monroe (tmonroe@worldbank.org). 18 BIG DATA in ACTION for GOVRNMENT