instability (Goldstone et al., 2010) or other events that may threaten liberal democracies ( Anderson, 2010), and anticipating social or political change through cyber-empowered political movements, social disruptions, or cultural conflicts (Bothos et al., 2010; Paris et al., 2010; Weinberger, 2011; Figure 3.6). However, rigorous methods for forecasting social patterns and social changes have not yet been fully developed.
Knowledge and Skills
Robust forecasting methods build on a solid understanding of the composition and structure of a system and the embedded interactions among system components and between the system and its environment (Boretos, 2011). Geospatial forecasting requires both deep domain knowledge and advanced skills in spatiotemporal analysis, modeling, and synthesis. Examples include regression statistics, spatial and temporal interpolation techniques, space-time prisms and trajectory models, cellular automata and agent-based modeling, artificial neural networks, evolutionary and genetic algorithms, computer simulation and ensemble techniques, and scenario-based planning that anticipates multiple possibilities.
Forecasts in the context of geospatial intelligence need to integrate both geospatial processes and domain processes to reveal patterns, relationships, and mechanisms that drive state changes. For example, activity-based intelligence—the predictive analysis of the activity and transactions associated with an entity, population, or area of interest—depends on an understanding of environmental, social, and cultural factors; individual space-time behaviors; and the spatiosocial processes that move and regulate activities of groups and the society.
New methods and analytical tools emerging from the computational social sciences are changing the education and skills needed for geospatial intelligence forecasting. For example, new approaches are being developed to address the validation and calibration challenges of agent-based and other complex systems models. Tools such as the Integrated Crisis Early Warning System have been developed to predict political events such as insurgency, civil war, coups, or invasion. The increase in volunteered geographic information and geotagged images or communications brings the field a step closer to short-term and near-real-time forecasts of event progression, such as the spread of wildfire or disease, or of social dynamics, such as perception or activities planning.
Education and Professional Preparation Programs
No university programs offer degrees in forecasting, and many science-based or business-based curricula emphasize modeling instead of forecasting. Courses in advanced methods for spatial and domain-specific processes are taught at senior undergraduate or graduate levels in a wide range of disciplines, including statistics, computer science, information science, electrical engineering, civil engineering, meteorology, geography, economics, ecology, criminology, epidemiology, and urban and regional planning. Geospatial forecasting requires an integrative treatment of spatial and temporal data and is still considered an advanced, specialized area of research. The few advanced spatial modeling courses available are commonly tailored to the faculty’s research interest, rather than providing a comprehensive coverage of analytical and modeling techniques. Examples of universities with strong programs in agent-based modeling include Carnegie Mellon University, George Mason University, and the University of Michigan (see Table A.10 in Appendix A). The Massachusetts Institute of Technology has a strong program in system dynamics.
Time-series analysis is the foundation for forecasting, and relevant courses are commonly taught in meteorology, geography, geology, ecology, economics, political science, and other departments that emphasize modeling and projections. Students learn how to detect temporal trends and to project them into the future using techniques such as harmonic analysis, wavelet analysis, and historical event modeling. Examples of programs that offer courses in these areas include the University of Oklahoma and the University of Washington (meteorology); the University of California, Santa Barbara, and the State University of New York at Buffalo (geography); and Harvard University and Princeton University (economics and political science; see Table A.10 in Appendix A).
Space presents another important dimension of forecasts. In human geography, spatial diffusion theory, central place theory, and time geography offer