The Data Science in Humanitarianism project studies how agencies such as the UN make use of data science to support decision-making and resource allocation in humanitarian and development work. Across the world, international institutions, national agencies, and civil society organisations are embracing “digital humanitarianism”. They are assembling and analysing vast digital data streams and mobilising online communities to use “real time” data to try to predict and respond to humanitarian crises, address unmet socio-economic needs, and allocate development assistance for maximum effectiveness (Eagle & Green 2014; Meier 2015; Hilbert 2016).
This project addresses the problem of how - or if it is possible - to distribute humanitarian aid and target development assistance using data science without undermining the integrity of those distributive decisions. It addresses, too, the problem of how to reconcile this use of data science with longstanding legal and policy principles, or adapt those principles to this changing practice, to minimise adverse effects. For example, concepts like "participation" and "the public" take on potentially radically different meanings to those traditionally embraced by policy makers and actors in the development and humanitarian spheres. Tackling these problems is important because of the growing prevalence of data science in humanitarian and development work, significant donor investment in such data science initiatives, and the lack of law and policy research surrounding the changing decision-making practices emergent in this context.