Data for policy 2021
The sixth international Data for Policy conference ‘Lessons for data-policy interactions after Covid-19’ ran online from 12th – 14th September. It hosted over 90 speakers and delegates from around the world.
As ever, presentations were broad-ranging and cross-disciplinary. The six thematic areas of interest of the standard tracks ensured that content extended from the highly technical through to real-world application. The three special tracks showcased highly topical work on data-driven innovation for sustainability, mobility as a service, and tools to tackle misinformation. Our dedicated social media team’s posts ensured a lively conversation throughout the conference https://twitter.com/dataforpolicy #dataforpolicy2021
The conference opened with a plenary panel chaired by Stefaan Verhulst (GovLab and NYU). The panellists Chinwe Ochu (Nigeria Centre for Disease Control), Brennan Lake (Cuebiq) and Chris Wiggins (Columbia University and New York Times) reflected on the impact of the pandemic and the importance of data in the response of their respective fields: public health, the private sector, academia and the media.
The two excellent keynote lectures were livestreamed on YouTube. Professor Karen Yeung (Birmingham Law School) focussed on facial recognition testing for her critique of ‘in the wild’ testing and experimentation with data-driven technology. Professor Dame Wendy Hall (University of Southampton) set out her vision for a Social Data Foundation for supporting multi-party data sharing, whilst maintaining the highest standards for data and governance and security.
Participants in the Conference commented on the broad array of subjects, engaging content and discussion, and the good organisation and time-keeping. Nevertheless, many were missing the interactivity of an in-person conference.
Zeynep Engin - UCL / Data for Policy
C. Leigh Anderson - University of Washington, USA
Emanuele Baldacci - European Commission
Jon Crowcroft - University of Cambridge & Alan Turing Institute
Andrew Hyde - Data & Policy Commissioning Editor
Innar Liiv - Tallinn University of Technology, Estonia
Christoph Lütge - Technical University of Munich, Germany
Barbara Ubaldi - OECD
Stefaan Verhulst - GovLab, New York University, USA
Masaru Yarime - Hong Kong University of Science and Technology
Laura Acion - University of Buenos Aires, Argentina
Omar Asensio - Georgia Tech, US
Feras Batarseh - Virginia Tech, USA
Eleonore Fournier-Tombs - University of Ottawa and World Bank, Canada
Giz Gulnerman - Committee coordinator
Anushri Gupta - Coventry University, UK
Shan Jiang - Tufts University, US
Joanna Kulesza - University of Lodz, Poland
Tian Lan - UCL, UK
Xiao Liu - McGill University / World Economic Forum, Canada
Lauren Maffeo - Steampunk, US
Keegan McBride - Hertie School, Germany
Alexander Monea - Geroge Mason University, US
Catherine Moore - Committee coordinator
Francesco Mureddu - Lisbon Council, Belgium
Jaron Porciello - Cornell University, US
Robby Cobby Avaria - Data & Policy Communications Editor
Zeynep Engin - Director
Emily Gardner - Community Manager
Andrew Hyde Data & Policy Commissioning Editor
Itzelle Medina Perea - Data for Policy Communications Editor
Call for Papers General Information
This standard track focuses on the high-level vision for philosophy, ideation, formulation and implementation of new approaches leading to paradigm shifts, innovation and efficiency gains in collective decision making processes. Topics may include:
- Data-driven innovation in public, private and voluntary sector governance and policy-making at all levels (international; national and local): applications for real-time management, future planning, and rethinking/reframing governance and policy-making in the digital era;
- Data and evidence-based policy-making;
- Government-private sector-citizen interactions: data and digital power dynamics, asymmetry of information; democracy, public opinion and deliberation; citizen services;
- Interactions between human, institutional and algorithmic decision-making processes, psychology and behaviour of decision-making;
- Global policy-making: global existential debates on utilizing data-driven innovation with impact beyond individual institutions and states;
- Socio-technical and cyber-physical systems, and their policy and governance implications.
The remaining categories represent more specifically the current applications, methodologies, strategies which underpin the broad aims of Data for Policy’s vision:
This track is concerned with data in its variety of forms and sources, and infrastructure and methods for its utilisation in policy and governance:
- Data sources: Personal and proprietary data, administrative data and official statistics, open and public data, organic vs designed data, sensory and mobile data, digital footprints, crowdsourced data, and other relevant data;
- Technologies: Artificial Intelligence, Blockchain, Internet of Things, Platform Technologies, Digital Twins, Visualisation and User Interaction Technologies, data and analytics infrastructures, cloud and mobile technologies;
- Methodologies and Analytics: Mathematical and Statistical models, Computational Statistics, Machine Learning, Edge Analytics, Federated Learning, theory and data-driven knowledge generation, multiple disciplinary methodologies, real-time and historical data processing, geospatial analysis, simulation, gaps in theory and practice.
This track focusses on governance practices and management issues involved in implementation of data-driven solutions:
- Data and algorithm design principles and accountability
- Local, national and international governance models and frameworks for data and associated technologies;
- Data and algorithms in the law;
- General Data Protection Regulation (GDPR) and other regulatory frameworks;
- Data intermediaries, trusts and collaboratives;
- Meta-data, interoperability and standards;
- Data ownership, provenance, sharing, supply chains, linkage, curation and expiration;
- Data sovereignty and data spaces.
This track examines the issues which must be considered in technology design and assessment:
- Digital Ethics: Data, algorithms, models and dynamic interactions between them
- Digital trust, and human-data-machine interactions in policy context
- Responsible technology design and assessment
- Privacy and data sharing
- Digital identification, personhood, and services
- Uncertainties, bias, and imperfections in data and data-driven systems
- Algorithmic behaviour: equity and fairness, transparency and explainability, accountability, and interpretability
- Human-machine collaboration in strategic decision making and algorithm agency
- Human control, rights, democratic values, and self-determination.
The following are areas which fall within the above categories, but are highlighted as being of special interest:
- Data-driven insights in governance decision making, black-box processing;
- Algorithm agency along with human and institutional decision-making processes;
- Government automation: citizen service delivery, supporting civil servants, managing national public records and physical infrastructure, statutes and compliance, and public policy development;
- Algorithmic ‘good’ governance: participation, consensus orientation, accountability, transparency, responsiveness, effectiveness and efficiency, equity and inclusiveness, and the rule of law.
- Human existence and the planet;
- International collaboration for global risk management and disaster recovery;
- Global health, emergency response, Covid-19 and pandemics;
- Sustainable development, climate change and the environment;
- Humanitarian data science, international migration, gender-based issues and racial justice;
- International competition and cultures of digital transformation.
Jaron Porciello, Cornell University (corresponding author)
Ulrike Hahn, Birkbeck College
Stephan Lewandowsky, University of Bristol
The COVID crisis has turned the world upside down. It has revealed societies’ fissures and pressure points as it has mercilessly revealed any lurking weaknesses in our existing systems and structures. The public and scientists have witnessed an explosion of scientific research across all disciplines –much it of understanding the nature of the virus itself—in addition to a well-spring of data science, meta-science and science communication, some of it drawing on state-of-the-art AI and machine learning tools designed to help scientists and non-scientists keep current on the explosion of knowledge.
The pandemic has brought into sharp focus questions surrounding the development, discussion, and diffusion of research. The wider issues they raise as they pertain to the ways science is and could be conducted in online information environments, whether this is among scientists themselves, in the interaction between scientists and policy-makers, or in interaction with the general public.
This special track will consider what we have learned as we emerge from the COVID-19 pandemic. What are the tools, systems, data governance models and types of experts that we need to foster science and help maximize its societal benefits well beyond the pandemic context? We will pay special attention to the role of media in the dissemination of new scientific findings alongside misinformation: expediency, if nothing else, during the pandemic has necessitated the use of extant social media platforms for science-to-science, science-to-policy, and science-to-public discourse.
We will bring together a mix of contributed papers and panel discussions to explore the relationship between discourse quality and algorithmic mediation. More specifically, we invite contributions on the following topics:
- COVID-19 required urgency to produce, react and make decisions based on scientific data. How have platforms tuned to the maximization of advertising revenue failed and succeeded to serve the epistemic goals of scientists? How did the fine-tuning of algorithmic behaviours on platforms like Twitter and YouTube end up impacting decision-making processes, psychology and behaviour of the individuals using these platforms? Are there new opportunities to re-purpose and extend existing tools to promote high-quality science discourse?
- Preliminary scientific findings find themselves in a complex set of dynamics between science, policy makers, public opinion, non-traditional `outlets’ for scientific research and the media that seem deeply problematic for the integrity of the scientific process. Many of the traditional systems of peer-review and science policy were arguably already challenged by the pre-crisis state of affairs, and the pandemic has placed huge additional demands on these system. Platforms that could reward expediency and transparency, in part by increasing usage of existing machine-learning algorithms, saw dramatic increases in usage and content. They have replaced some quality-control and review functions that were previously designed for humans. If the principles of expediency, transparency and informal review are to become part of the new normal, then what models of data governance and science policy do we need to encourage? What types of sociotechnical models do we need to encourage unbiased decision-making?
- The pandemic clearly challenged the traditional science-to-policy interface. Scientists in all disciplines were called on to act as intermediaries and build trust between the general public and policy-makers. Scientists were vetting “emergent science” as it was published on social media platforms, resulting in at least one high-profile retraction. This has pushed the job of technical discussions and scientific peer-review into a new social space and one where publishers do not wield as much power over the dissemination of scientific ideas. However, oversight is important, and in this panel we invite novel ideas for governance systems, online tools and financing mechanisms that we need in order to make successful and long-lasting changes towards creating new systems for science.
- There is evidence that shows that the same misformation tactics and campaigns that have been used for climate change (among other issues) are the same ones casting aspersions about the COVID-19 vaccines. How can we prevent bad actors from overwhelming public systems and reduce opportunities for misinformation campaigns based on expedient science now, and in the future?
Masaru Yarime, Hong Kong University of Science and Technology
Data-driven innovation, including the Internet of Things (IoT), blockchain, and artificial intelligence (AI), has significant potential to address various challenges identified in the Sustainable Development Goals (SDGs). Smart energy grids based on blockchain make it easier to integrate renewable energy sources and balance energy supply and demand smoothly, improving energy efficiency and reducing carbon dioxide emissions and air pollution. Smart infrastructure monitoring systems equipped with IoT sensors enable us to measure weather conditions precisely and strengthen urban resilience to natural disasters such as floods and typhoons. AI-based medical devices support conducting diagnosis and treatment efficiently, providing accessible and inclusive health services to all.
As environmental, economic, and social aspects are increasingly interwound, various kinds of data that are getting available from a variety of sources through sophisticated equipment and devices need to be deployed effectively to tackle multifaceted sustainability goals and targets. It is hence critical to collect, share, and use relevant data through cooperation and collaboration among stakeholders. The open data policy will facilitate data disclosure and exchange, contributing to creating innovation.
There are many policy challenges, however, that we need to consider in properly utilizing data for innovation for sustainability. Technical issues related to data, such as metadata tagging, quality control, cleaning and error elimination, and interoperability between various standards, must be addressed to support data sharing. Stakeholders have different interests and motivations and would not necessarily be willing to disclose or exchange data with each other, which would require us to consider a proper balance between open and proprietary data. Serious concerns are also raised about dealing with sensitive data in terms of security and privacy. Micro-targeted nudging for sustainable behaviour based on detailed personal data could involve manipulation or paternalism.
Various policy approaches can be considered for data governance. The government can be in charge of governing public data, whereas platform enterprises in the private sector also play a critical role in assembling and managing an increasing amount of data. A data trust can be established as an independent institution to make decisions about who has access to data under what conditions, how that data is used and shared for what purposes, and who can benefit from it. Transparency and citizens’ participation and engagement in the processes of data governance are particularly emphasized in encouraging social acceptance and inclusiveness in pursuing sustainable objectives.
Data-driven innovation poses a difficult challenge to policymaking. The speed of technological change is rapid, and the path of its evolution is not entirely predictable or explainable. That would produce a widening gap between technological change and institutional readiness. Also, various sectors, including energy, transportation, and health, which were not connected previously, are increasingly integrated through data in cyber-physical systems. Novel policy approaches, such as regulatory sandboxes, would be required to incorporate the ability to learn from real-world use and experience and improve performance through adaptation.
In-depth research needs to investigate what kinds of policy frameworks and measures would be effective in collecting, sharing, and using data among stakeholders and what impacts would be made on facilitating data-driven innovation while addressing societal concerns, including data security and privacy. This special track invites theoretical as well as empirical studies that examine various policy measures and approaches to facilitating data-driven innovation and addressing key issues involved, such as the ownership of and accessibility to data, interoperability and integration of data, incentives to the collection, disclosure, and sharing of data, the protection of data security and privacy, and trust and engagement in data governance.
Possible questions we would discuss in this special track include, but not limited to, the following:
- How are various kinds of data collected, shared, and used for innovation among stakeholders?
- What incentives are provided to stakeholders with different interests and motivations to facilitate data sharing?
- What kinds of governance systems are established to manage data availability, accessibility, and ownership?
- What policy measures and institutional arrangements are introduced to deal with sensitive data in terms of data security and privacy?
- What are the impacts and consequences of policy measures on facilitating innovation while addressing societal concerns?
Case studies in different countries and regions are particularly welcome to examine the mechanisms and processes involved in data collection, sharing, and use for innovation, as local specificities of the relevant actors and institutions would be significant. Policy implications and recommendations are explored for maximizing the potential of data-driven innovation while minimizing risks to individuals and communities in addressing SDGs.
Dr. Ronit Purian and Avi Cohen, SYN-RG-Ai, Tel Aviv University, and CODATA
Participants in this track will present approaches and methods to better understand urban dynamics, identities and spatial behaviors that incorporate collective actions in cities today. Specifically, willingness to share data in decentralized systems through careful design is at focus, assembling data trusts for communities – in different ecosystems and social groups – while reconstructing trust in government and in the nation state and institutional practices.
Personal data sharing for the benefit of society at large is a goal inspired by the Covid-19 contact tracing applications. The voluntary use aimed at utilizing the value of crowdsourcing and self-reporting, to appropriate the very same principles of citizen science; however, acceptance rates were low (similarly, institutional distrust is so widespread that vaccination is at risk, even if provided voluntarily and free of charge). We wish to better understand the reason for this failure in terms of service design and institutional practices, i.e., what makes a trustworthy technological and organizational design.
On the continuum – between the elementary disclosure of information, to beliefs creation and trust building – lays a level in which a contract is established between the citizen and the dedicated authority. The contract level is core to technology use. Contract violation decreases users intention to reuse the system, and therefore, violating this psychological contract with the public is a moral hazard. Citizens were suspicious of the state-controlled applications, although government intrusion into their personal life with applications that applied a decentralized architecture was implausible. Is that possible that, rather than mechanisms to avoid privacy invasion, other design aspects should have been considered and emphasized? Dull interface and functionality could explain the evident failure of contact tracing applications. How can we identity design aspects and institutional practices that alienate participants?
In this track, we wish to extend our knowledge of information systems (IS) design. How can system use elicit empowerment, engagement and trust building? This question refers to data intermediaries, informed consent, data ownership, modularity (to be able to switch between decoupled apps and data servers, not losing data due to vendor lock-in) and the ability to easily reuse existing data from other apps – as proposed by SOLID, and more. In addition, the question refers to a series of new socioeconomic contingencies, mainly: the chronic-rural poverty – and the new-urban poor.
To illustrate this, the World Bank (2021) presents the “multidimensional character” of poverty in rural areas, where low levels of educational attainment are common to both poor and nonpoor; but at the same time, telework may challenge the education advantage of cities and the productivity effect of agglomeration in dense urban areas. Early results from rapid-response phone surveys show that a large share of the new poor will be urban. Moreover, “many of the new poor are likely to live in congested urban settings and to work in the sectors most affected by lockdowns and mobility restrictions; many are engaged in informal services and not reached by existing social safety nets” (World Bank, 2021; p. 143). Thus, the character of poverty is “multidimensional” also in urban areas, and Covid-19 is likely to have distinctive effects on poor people who are urban residents. An inclusive and sustainable mobility service should consider, accordingly, specific populations (e.g., undocumented work immigrants), point of departure (who live in congested informal settlements), and destination (who work in the informal sector). What should be the focus of data interoperability in such mobility services?
We believe that identifying the most vulnerable groups in society is important, not only to eliminate health disparities and adverse health effects related to climate variability and societal gaps, but to increase resilience in the sense of systemic engagement, caring about ecological and social systems, nurturing mutual responsibility and a sense of individual ownership.
The climate crisis, and health and economic problems are intertwined through our urban-digital life, and call for an urgent systemic transformation. Therefore, our overall aim is to further develop the systems and the societal practices that encourage healthier and more sustainable spatial behaviors.
We plan to carry out a collaborative interdisciplinary track that will define focus areas, activities, and priorities, e.g.: Needs assessment and knowledge gap analysis (the desirable data sources and tools); reflections on current studies (case studies – challenges, solutions; and how to generalize); operational research plans (national surveys, public concerns and perceptions in questionnaire items), and the like.
From the domain viewpoint: to define a cluster of mobility-related problems that are encountered by many cities.
From the data viewpoint: to define the data components we share in our research. The goal will be to facilitate open data, knowledge sharing, and data integration.
From the service viewpoint: to define the data sources and tools we use; what are the preferred integration and granularity (e.g., local-global); and how to achieve that.
To address such goals, we wish to invite researchers from different disciplines, who are interested in: Urban Resilience, spatial behavior, flexible mobility, data ownership, service design, and any other topic within the broad domains of smart cities, information systems, data science, and behavioral sciences.
The track will integrate the following aspects:
- Accountability and privacy by design, security, open code
- Beliefs creation and trust building, attitudes, communicative action
- Civic mindedness in global cities
- Data integration and interoperability
- Global citizens and work immigrants
- Human-machine interface and rich system use
- Mechanism design, moral hazard, PPP, P2P
- Smart cities and urban networks
 SOLID: https://solid.mit.edu
 World Bank (2021). Global Economic Prospects, January 2021. Washington, DC: World Bank. http://hdl.handle.net/10986/34710 and https://openknowledge.worldbank.org/handle/10986/34710
Review and Assessment process
for Conference submissions
All submissions to the Conference were assessed by peer review for their suitability for the Conference, according to the following criteria:
- Potential contribution to the debates in the field
- Potential for stimulating debate in the Conference
- Freshness of the content, novelty and originality
- Formulation of the research/policy question
- Data and methodology
- Quality of writing and presentation
In addition to assessment of suitability for the conference, full papers received peer review aligned to that of the Data and Policy journal .
POST-ACCEPTANCE OF CONTRIBUTION
All presenters were required to submit a video of their presentation before the conference. This was shared with all registered attendees via the Data for Policy YouTube channel.
All authors, including those invited to submit papers to Data & Policy, were encouraged to upload a discussion paper to the Data for Policy community page on Zenodo.
Data for Policy is a fee-paying conference.
For all questions relating the conference, please contact firstname.lastname@example.org; and for questions relating to Data & Policy, please contact email@example.com