Tracking the data revolution: responsible re-use of data

<aside> 🌎 The Global Data Barometer will provide a biennial snapshot of the responsible re-use of data across the world

</aside>

Over recent decades, questions around data have moved to the heart of political discourse. Private firm, charities and governments around the world are increasingly deploying data to drive their decision making, raising critical questions about who is included or excluded, helped or harmed, by these practices. Flows of data are shaping urban life, public service provision, and responses to major policy challenges such as addressing climate change, tackling corruption and improving health care provision. Companies and countries across the world are competing to be data leaders, particularly in the application of artificial intelligence, identifying this as critical to their future economic development.

All in all, data has come to be seen as a powerful resource: an asset that can be used for private gain, or to solve shared problems in new ways. At the same time, threats to privacy, biases embedded in data and its analysis, and inequalities of power created between data haves and have-nots, have also come to be seen as new risks to be managed.

After a decade of focus on open data, discussions in the open government space have broadened to look at ways to generate the greatest social value from data resources, whilst managing data risks. We describe this as responsible re-use of data.

A focus on responsible re-use starts from the premise that good quality data, and the capabilities to use it, are important tools of sustainable development, and that the value in data should be shared widely. But it also recognises that, dataset-to-dataset, and sector-to-sector, there are different dangers to be aware of, and that governments have a critical role in creating an environment where the power of data is used for social good, and is not used to create harm.

A new global study: bridging the gap

<aside> 🌎 The Global Data Barometer is built around identified user needs for data, evidence and ongoing assessment. (Read initial stakeholder engagement summary here)

</aside>

The design of the Global Data Barometer responds to critical knowledge gaps in how policies and practices of responsible re-use are unfolding across the globe. It will focus on:

Continued tracking of open data. Governments and civil society advocates have a continued need to benchmark their open data activities. The GDB will provide a comparable time-series with the predecessor Open Data Barometer, whilst addressing limitations of the earlier study. This will support country level assessment and action, and maintain a global picture of how far open data has become embedded, or how far initiatives have stalled.
Going beyond open data alone: internal, shared and collaborative data use. As the discourse around data has become more sophisticated, looking at a spectrum from closed, to shared, and then open data, measurement tools have not kept up: giving only a partial picture of the landscape. The GDB will address the relationship between capacity and use of data inside government, the positive and negative aspects of data-sharing arrangements, and the growth of data collaboratives and other models of public and private sector collaboration.
Addressing data literacy and data science. The State of Open Data report pointed to major gaps in data literacy and capacity for use data, highlighting the need to better track the availability of data science skills and critical data literacy across various sectors. The GDB will use a range of methods to identify existing skills, and interventions in capacity building for countries to harness the data revolution.
Emerging practices around artificial Intelligence. The development and application of artificial intelligence relies heavily on access to data. It's vital to understand which countries are collecting and providing AI-ready datasets, and to track the steps are being taken to address issues of bias or under/over-representation of particular languages or populations in the training of machine-learning systems. Maintaining a focus on public data, the GDB will look at readiness to deploy AI in the use of data, the collection and sharing of data with AI-relevant features, and uses and impacts of data that involve AI. Through the global expert survey approach of the GDB, this will offer a unique dataset of AI use-cases from across the world, providing a much more representative view of activity that is captured by current studies.
A sectoral focus. Work to release open data, deploy algorithmic decision making, or build capacity for data use may take place at a whole-country level. But more often, it involves work in specific sectors: health, agriculture, justice, transit and so-on. Equally, initiatives to promote the creation of new data infrastructure, more openness, or use of data for social impact, tend to have a sectoral focus. By looking at readiness, implementation, use and impacts of data on a sector-by-sector basis, the new GDB will map the comparative advantage of different countries, and will support global initiatives to develop baselines, to refine their strategies, and to monitor their progress.

<aside> 💡 Setting the boundaries: a note on scope

Through processes of 'datification' almost every part of modern life can be viewed through a data lens.

The Global Data Barometer is centred on public data understood as data that is generated or used by entities carrying out a public task such as the delivery of public services, or the management of public resources. In most cases, this excludes personal data, understood as data that identifies individual people (for example, social media data) and corporate data that does not become involved in public tasks (for example, companies internal supply chain data, or data used to train computer vision models in industrial machines).

As such, in the initial editions of the Global Data Barometer we are not trying to answer questions about the extent, impact, resource or risks of social media platforms like Twitter and Facebook, nor to explore in any depth issues of data-driven personalisation of consumer services, or surveillance of consumers and citizens. The extent to which a country is making use of data or AI solely within industry falls outside the scope of the proposed study. Scientific data also falls outside our core focus.

In most cases, the data we are focussed on may are capable of being represented as structured text or quantitive tables or plotted on maps. Datasets covered by the study may be both small and big data. With a few exceptions, we are not focussed on image, video or audio data.

</aside>