A Research Collaboration between UN Global Pulse Lab in Jakarta and UNSW Sydney
The University of New South Wales Sydney and United Nations Global Pulse Lab in Jakarta are delighted to be working together on the ‘Data Science in Humanitarianism: Confronting Novel Law and Policy Challenges’ project. The project, funded by the Australian Research Council, seeks to understand how decision-making by international agencies and nation states may change as they make use of data science to support their humanitarian, risk reduction and development work.
Because data science transforms how and what we count, and how we make meaning of the world, data science may challenge what we have historically understood as a sound basis for law and policy action. Validation and modelling techniques that have underpinned much law and policy decision-making in the past may not readily transfer to analysis of large volumes of unstructured, digital data.
The overall goal of the project is to identify ways we might retool existing law and policy thinking, on the one hand, and help improve data science practice, on the other — in both respects to confront equity and legitimacy challenges that may arise from the digitisation of humanitarian work.
Team members from the UNSW have visited Pulse Lab Jakarta, spending time getting to know the team, developing their own understanding of tools the Lab has developed, and getting a glimpse of how the team works. Understanding the multiplicity of stakeholders that Pulse Lab Jakarta serves — including data donors, financial contributors which include international aid programmes, national governments, and of course end-of-the-line beneficiaries — reveals something of the enormous complexity of the ‘data for good’ movement.
Different actors have different interests and objectives, and their decisions are governed by different legal, ethical, social and epistemic (knowledge-making) regimes. More than that, the data provided by different actors — including administrative (government) records, commercially generated data, such as mobile network data, social media data and scientific data — all bring with them distinct imperatives and ways of making sense. The task of integrating them effectively for humanitarian ends poses significant and often unacknowledged challenges.
One example concerns the incommensurability of some statistical and data science methods. A statistician in a national statistics office will typically proceed first by determining what they want to know, modelling the information required, collecting the information, ensuring it is suitably representative and randomised, and then running calculations. Data scientists work with pre-existing data — typically, unstructured, messy data from a range of sources, often framed at a variety of scales — to try to generate insights that might support better decision-making.
Because of the enormous reach and cost saving-potential of digital data, and the possibilities they hold to yield timely and fine-grained insights, new techniques are urgently needed to operationalise a range of digital data types. These techniques must ensure the robustness of claims made about the world for humanitarian and development purposes and support the legitimacy of action taken (or not) and resources allocated (or not) on the basis of such claims. This presents, at once, a technical and a policy conundrum.
Pulse Lab Jakarta is in the middle of these contending ways of sensing, recording and developing meaningful insights. The team works through these challenges on a daily basis in its combined role as data-broker, eco-system enabler and data science lab. This position, at the heart of a process of transformation still unfolding, makes Pulse Lab Jakarta a wonderful nexus and collaborator for research.
Over the next three years, the legal and social science researchers in the UNSW team will continue conducting ethnographic research with the Pulse Lab Jakarta team, and with their collaborators, as they work to improve understanding and aid navigation of the rapidly changing decision-making and data environments in which they work. Modern governance architecture was, in large part, built around statistical thinking. We want to detect which parts of this governance architecture are likely to come under greatest pressure as underlying norms and practices shift, and how governance frameworks might adapt to that.
Meanwhile, the two computer scientist members of the UNSW team (one of whom has both statistics and computer science training) will be working on improving technical methods at the intersection of big data, the sustainable development goals and official statistics. We hope, through this work, to help bolster data science methods’ robustness, better grasp contingencies that arise from particular analytical practices, and support improved development and humanitarian practice.
Decision-making 2.0 for development and humanitarianism is very much a work in progress. Sustained, interdisciplinary, cross-sectoral collaboration of the kind in which Pulse Lab Jakarta and UNSW Sydney are engaged is invaluable if it is to advance responsibly.
Pulse Lab Jakarta is grateful for the generous support from the Government of Australia.