April 17, 2017

Harnessing Big Data for Development and Humanitarian Action

poster of the Sustainable Development Goals

In this interview with Robert Kirkpatrick, director of UN Global Pulse, he discusses the role of big data in achieving the Sustainable Development Goals and outlines three areas of impact. Global Pulse is an initiative of the United Nations that works to harness data science for sustainable development and humanitarian action. Prior to joining the UN, Robert cofounded and led software development for two pioneering private-sector humanitarian technology teams, first at tech startup Groove Networks, and later as lead architect for Microsoft Humanitarian Systems. From 2007–2009 he served as CTO of Google’s nonprofit health and disaster-tech spinoff InSTEDD. Last week's installment on this blog introduced the topic of data's role in the Sustainable Development Goals.

Q. During the member-state negotiations towards the High Level Political Forum last year, I’ve sensed a certain level of excitement among the member states about a “data revolution” and data-driven policy. What do you mean by data revolution?

Increasingly, everything people do produces data that may be used in myriad ways, for good or ill. The data revolution in its broadest sense, which typically involves the availability of big data, together with the emergence of artificial intelligence and the Internet of Things, is already transforming society in ways we do not yet understand. When we talk about big data for development and public policy, there are at least three revolutions at stake, three distinct transformative processes that will play significant roles in global efforts to achieve the Sustainable Development Goals (SDGs), and big data innovation figures prominently in each.

So the first “data revolution” is a measurement revolution. Here I mean getting to the finish line in having the timely, accurate, complete and high quality statistical data required to support effective policy planning and impact tracking in every part of the world. Big data here may not only provide opportunities to leapfrog directly to using new approaches to generate indicators, but also to do so in ways that increases frequency and reduces costs when compared to traditional approaches. In other cases, big data may be used to fill into periods between official censuses, tracking trends in key indicators and alerting decision makers to concerning trends in regions where they might wish to conduct surveys to understand more about what is happening. There are National Statistics Offices all over the world experimenting with big data to see where it may be used to generate reliable proxy measures of indicators, including many of those used to track SDG progress.

The second is an accountability revolution. Here we are talking about turning to these new sources of data as a continuously, automatically updated evidence base for understanding the effectiveness of policy, and for empowering communities to use data analytics in innovative ways to hold policymakers accountable for results. Today’s open government movement arose from the recognition of a power asymmetry between those who have data and those who do not. While there is great value in opening up government surveys and census data, how useful it is for helping ordinary citizens to address current challenges is often limited by how many years out of date it is. Imagine a world in which anyone can track the monthly, weekly, or even daily effects of a new policy. How rapidly are different communities recovering from that disaster? How many jobs has that new program created — and how many of those jobs were filled by women?

The third is a management revolution. When we look into the 2030 Agenda, the challenges arise not only from the inherent (and appropriate) complexity of the framework, but also from the fact that we must implement it in a world of accelerating change. Development can’t be viewed as an incremental linear process, and our linear incremental institutional processes and our log frames are insufficient to keep us ahead of the curve. Policymakers must move from a modus operandi in which they rely solely on using records of the past to plan the future, and then taking an occasional snapshot along the way to track progress, to one in which they have access to rich sources of real-time information and feed this information into predictive models that let them respond proactively to new dynamics and emerging risks.

In one sense, I suppose, we’re talking about development practice borrowing a page from humanitarians, who rely as much on information coming through their walkie-talkies to operate as they do on needs assessment forms, but it also means coming to rely on unfamiliar technologies to try to peer into the future at every turn.

Q. What role does Global Pulse play in the implementation of the global goals?

Implementation of the 2030 Agenda is the responsibility of national governments to their citizens. The UN’s role is to support that effort in various ways, including by anticipating their needs and serving as a catalyst for the many innovations required.

Our strategy around implementation involves advising national institutions directly, as well as providing both technical assistance and policy guidance to UN agencies, funds, and programs in their respective efforts to support national institutions. Global Pulse is one of the innovation initiatives currently operating within the UN. We provide joint innovation as a service through our Pulse Labs in Jakarta (Indonesia), Kampala (Uganda), and NYC. We work with and support innovation initiatives at UNICEF, WFP, UNDP, UNHCR, OCHA, WHO, etc.

Part of what we do at Global Pulse is focused on learning how data science and analytics may contribute to smarter operational decision-making along these lines, whether it’s about supporting the achievement of sustainable development, improving humanitarian action, tackling climate change, or building peace. Another part is about policy innovation around data privacy and responsible use, to make sure efforts to help don’t end up causing more harm than good, and helping UN agencies and national institutions adopt new tools and approaches. There’s a bit of advocacy work as well. Part of this outreach involves engaging priority industries that have relevant data, such as mobile, banking, retail/ecommerce, and postal operators around the opportunity to put it to work for the 2030 Agenda. For example, the mobile industry through the GSMA just announced a new “Data for Good” strategy around the SDGs, which we look forward to supporting. At the same time, we also see a need for more public engagement around the fact that using big data doesn’t mean compromising privacy, and that there is a daily opportunity cost of not having all of this rich information put to use to improve delivery of public services and strengthen accountability. 

Q. What is the scale of Global Pulse’s work?

We’re still relatively modest in terms of the work we do in SDG indicators. We’re just now seeing the work getting started. We are involved in a couple of projects that aim to support the official statistics community in SDG measurement. For example, we are partnering with the Ugandan Bureau of Statistics on creating poverty maps through remote sensing by counting roofs of buildings and automatically classifying the material as either thatch (less expensive) or corrugated aluminum (more expensive). We have a number of humanitarian projects underway, from using machine learning on AIS tracking data to recognize rescue events in the Mediterranean, to using GPS data from delivery trucks to optimize sanitation services for UNICEF in the Za’atari Refugee Camp in Jordan, to supporting UNOSAT in automatic labeling structures in satellite imagery, to leveraging our partnership with Twitter to help UNHCR monitor perceptions of refugees across Europe. Our hazegazer.org platform has been incorporated in the early warning system in the Office of the President of Indonesia to identify hotspots and monitor reports of the impacts of the fires in Sumatra. In Uganda, our Lab developed speech recognition for three Ugandan languages (more on the way) and is using it to turn public conversations on rural FM call-in radio shows into machine-readable text, making it possible to generate real-time insights from locations where few people ever access the internet. There are dozens of projects in the pipeline, and demand is growing. The privacy policy work is advancing as well: we have been very much involved in steering discussions of the UN Development Group’s newly adopted guidelines on big data privacy and ethics. Our first big data open innovation challenge, http://dataforclimateaction.org, just launched on March 9.

Q. How far are we in terms of getting the buy-ins for using big data from the leadership in different UN agencies and programs?

Right now, we are seeing more and more UN agencies interested in new ways of working, all the way up the chain of command, as well as cross-cutting internal communities, such as program evaluators. Also, the SDGs have had a catalytic effect in the market. Development used to be owned by corporate philanthropy and corporate social responsibility departments, but now the market is responding to the SDGs as a core business agenda. Along the way, they are seeing data philanthropy as a business model, identifying where it makes sense to give insights away for free (e.g., disaster response) and where there is an opportunity to do well by doing good (e.g., improving public transportation systems). So we have a lot more companies coming to us unbidden, asking how they can put their data to work to support the work of the UN.

Q. The estimated total cost of getting the data needed for SDG measurement is known to be $5 trillion per annum. Is there any role that big data play in tackling this issue?

I do believe there are significant opportunities here to revolutionize monitoring of development progress, though it will take a few more years before we really start to see these approaches getting officially adopted. At Global Pulse, the use cases we pursue aim to help us learn how we can use big data to assist countries in measuring and achieving the SDGs. We spent years conducting research and producing proofs of concept. These days, we and our partners are laser focused on getting enough usable tools into the hands of real users to understand the costs and benefits of incorporating analytics into program implementation. We need to understand the training, licensing, and cloud computing requirements, the value-added vs. traditional approaches, and the barriers to scale.

The views expressed in the Government Innovators Network blog are those of the individual author(s) and do not necessarily reflect those of the Ash Center for Democratic Governance and Innovation, the John F. Kennedy School of Government, or of Harvard University.

Related Topics

Related Topics