Introduction to Spatial Agent-Based Models


Cross_scale_dataCourse Summary: This 5-day short course will serve as an introduction to the theory and practice of spatially-explicit agent-based modeling (ABM). You will learn the essential theoretical background and technical expertise needed to conceptualize, build, and analyze your first ABM. This course will guide you through the basic phases of the ABM research process: formulating a research question, specifying a model, creating a simulation and interpreting the output. The course combines lectures with hands-on model-building sessions where you will build a model using NetLogo to acquire basic and intermediate programming skills. More advanced students are welcome to build a model in a programming language of their choice. This will be an intensive, week-long immersion in ABM concepts and methods with reading and short writing assignments each day, and a ‘final project’ consisting of a simple model and standardized documentation to be published in the OpenABM (www.openabm.org) repository.

Target Audience: This course is intended as a foundational course for anyone interested in adding ABM to their analytical toolkit, regardless of prior modeling experience. Much of the course material and lessons will emphasize spatial ABM for understanding the dynamics and interdependencies of humans and natural systems (i.e., socio-environmental (S-E) systems). Applicants whose research or teaching focus on such topics will be given preference, but applicants with other areas of interest are also welcome. The course material will be structured for students with little to no experience with agent-based modeling and/or programming, but it could also be of interest to researchers/faculty with limited agent-based modeling experience. Target class size is 12-15, so space is limited.


ABM Theory

  • Essential background: What are S-E systems, and why are ABMs one of the best tools for understanding them?
  • Building blocks of spatial processes: forces of attraction and segregation, individual mobile entities, and processes of spread.
  • Building blocks of computational social science: objective functions, decision-making theories and models, social networks, and agent learning.

Model Development

  • Designing and building an ABM: model development objectives and best practices, including using the standardized Overview, Design concepts, and Details (ODD) protocol for model documentation.
  • Code version control with git repositories
  • Interfacing models with spatial databases, harmonizing spatial and non-spatial data for model parameterization.

Model Analysis

  • Concepts and methods for model evaluation: explanation versus prediction, multifinality and equifinality, outcome and structural accuracy, pattern-oriented modeling.

For more information, please see the course page on the SESYNC website here.