Big Data for the Prediction of Forced Displacement

Abstract

In a world where every 80th person is now forcibly displaced, using big data sources to improve planning processes is no longer a question of if, but a question of how. In recent years, UNHCR has intensified its efforts to integrate a variety of data sources, ranging from satellite imagery to newspapers to online digital data, into estimates of refugees and persons of concern. These novel data sources offer UNHCR an opportunity to improve planning about early warning and acute crisis situations. This paper outlines the potential of big data for practitioners within the area of predictive work in the humanitarian sector and presents examples of how some of those data sources are currently used in the organisation. In particular, it considers the opportunities and challenges of aligning big data with UNHCR's goals of accuracy, scalability and sampling bias adjustments.