Mesa, agent-based modeling in Python

Abstract

The Mesa framework is the predominant tool for agent-based modeling and simulation in Python. Mesa, created in 2014, has a GIS extension called Mesa-Geo, added in 2017. It has received substantial attention and contributions from the open-source community. This community effort has led to its applications in a diverse range of research, with more than 350+ published research papers utilizing the Mesa framework. In this talk, we will present how researchers use and think about space in their published research. We will then dive into a brief overview of spatial abstractions in agent-based modeling, from the most simplified space to a realistic representation of the world using Mesa-Geo. Through this process, we will cover how Mesa and Mesa-Geo models are constructed. Lastly, we will cover open-source spatial additions by researchers using Mesa, such as interior building space and synthetic populations.