Data is the lifeblood of decision-making, but how should public administrators navigate the digital deluge and who should be joining them on the journey? Can Big Data augment official statistics, or will it always be undermined by its coverage errors and biases? What about the individuals and groups in the shadow of the digital divide and those in the blind spot of official statistics, can we realise the ambition of the 2030 Agenda to leave no one behind?
With these questions in mind, the UN Development Group for Asia-Pacific recently organised a three-day data workshop for Resident Coordinators and Heads of Agencies from across the Asia Pacific region in collaboration with UN Regional Thematic Working Group on Statistics and Pulse Lab Jakarta. It’s hard to capture the full wonder of such an event in a blog, but we’re going to give it a go.
The waltz of big data and official statistics
When it comes to national data ecosystems, there were insightful presentations from UN country teams on system mapping and the utility of existing data, and from UN ESCAP on the role of national statistics offices and the production of official statistics, in relation to monitoring the Sustainable Development Goals at both a global and national level.
The census was reaffirmed as the backbone of national statistical systems. The workshop participants took this on board and several commented on the need to provide more political attention and support to the production of the census. In terms of the connection to Big Data, it was highlighted how the census is used to calibrate insights by weighting the values associated with underrepresented groups.
The connection between the census and Big Data does not stop at post-stratification methods. The discussions also touched on how some forms of Big Data are contributing to the census and other official statistics. We heard of inspiring examples from Scandinavian countries of how the population and housing census is compiled exclusively from administrative data (no more questionnaires!). For those countries where it is not safe to assume that every resident interacts with the state at some point during the census cycle, we heard a nice example from Afghanistan where survey data and satellite imagery were combined to create high-resolution population estimates. But very vulnerable groups will continue to require targeted research initiatives, there are no shortcuts in this area.
Moving forward, it is not a case of Big Data or traditional official statistics. It is a symbiotic relationship that will grow as new methods and data collaboratives are developed.
The 2030 Agenda requires a step change in data collaboration
It goes without saying that we are only going to achieve the 2030 Agenda if we embrace a more collaborative and partnerships-driven approach to development. Digital disruption, whilst causing transformative change across the world, is just scratching the surface of the development sector. And yet mobile networks and other forms of connectivity can enable access to information and services like never before. We need to view these networks as platforms upon which we can deliver development, as well as partners through the lens of shared value.
In this respect, it was great to hear from Digicel, Twitter and DataKind on how they perceive their role in releasing shared public-private value, how development is changing and what we all need to do to stay relevant to the people we serve. It was also good to learn from Flowminder, LIRNEasia, J-PALand Empatika on their cutting edge research projects, and how non-traditional data sets have the potential to contribute to social good. ‘Data Collaboratives’ discussed by UNICEF was also a useful example of bringing in the private sector and academia to tackle complex socio-economic problems.
Tapping in to Big Data, citizen generated data and open data sources can be more useful in understanding perceptions of a population than traditional surveys. As a colleague from the UN team in India analogised, these data sources give us a fuzzy understanding of what’s happening around the car, right now, which is useful alongside the clear image in the rear-view mirror.
We need to embrace the change within the UN, so we can in turn help those who need us to be at the forefront of development innovation.
Doing data for good, and doing it ethically
Given the growing demand for using Big Data sources for programming in development and humanitarian action, it’s vital to understand the privacy implications of the projects we are about to embark on to ensure that we understand the risks and potential harms we may come across.
The workshop included an enlightening session from UN Global Pulse’s legal supremo on data privacy and ethics, followed by a robust debate. It was great to hear the participants commit to research that is ethically sound, not just legally compliant. In addition, the debate around informed consent was fabulous. When is consent truly informed, and to put it nicely, when is it just consent to “cover the researcher’s derrière”? The debate rages.
The UNDG Guidance Note on Big Data for Achievement of the 2030 Agenda was highlighted as a great starting point for UN employees wanting to work with Big Data, as was the data innovation risk assessment tool.
Due to the timing of the workshop, the EU’s General Data Protection Regulation was also a hot topic of debate, and the key message from the room was that the UN needs to engage further in a normative role in helping member states to address risks to rights, as well as risks from bias in algorithmic decision-making and the perpetuation of inequality.
The politics of data
Data, information and insights are political, and we need to think through the political dimensions of the stories being told by data. Gaining access to data can be difficult particularly for business-sensitive data, but transformative SDG data in the Asia-Pacific region should be a priority.
It is good to remember that data are not just numbers, and we were reminded to check our assumptions by Empatika during the workshop. So many assumptions do not hold up to scrutiny, so remember to check in with the people behind the numbers. Quant + Qual is a great formula for success, and not always in that order.
Building a culture of collaboration and partnerships to ensure that the work that we do addresses the needs of the most vulnerable and marginalized is key. The UN data strategies at the level of UN country teams can ensure that we understand and wisely interpret the data at hand, know about data gaps and draw on internal and external resources and capacity to fill data gaps when required.
As the UNFPA Executive Director said in a special address to the event: “the moon moves slowly, but covers the town.” Have faith in ourselves, and in others.