From earth observation to human observation: Geocomputation for social science

It is possible to obtain vast amounts of spatiotemporal data related to human activities to support the study of human behavior and social evolution. In this context, geography, with the human–nature relationship as its core, is undergoing a transition from strictly earth observations to the observation of human activities. Geocomputation for social science is one manifestation thereof. Geocomputation for social science is an interdisciplinary approach combining remote sensing techniques, social science, and big data computation. Driven by the availability of spatially and temporally expansive big data, geocomputation for social science uses spatiotemporal statistical analyses to detect and analyze the interactions between human behavior, the natural environment, and social activities; Remote sensing (RS) observations are used as primary data. Geocomputation for social science can be used to investigate major social issues and to assess the impact of major natural and societal events, and will surely be an area of focused development in geography in the near future. We briefly review the background of geocomputation in the social sciences, discuss its definition and disciplinary characteristics, and highlight the main research foci. Several key technologies and applications are also illustrated with relevant case studies of the Syrian Civil War, typhoon transits, and traffic patterns.