How Useful are Urban Models in 2022?
“All models are wrong, but some are useful” is a widely used aphorism attributed to statistician George Box. This article focuses on exploring how useful urban models have become in 2022 for answering questions like where, when, and how urban areas will develop, how infrastructure and development regulations will impact those outcomes, how alternative strategies can help reduce greenhouse gas emissions or reduce the housing affordability crisis, and how infrastructure and development regulations can be more effectively coordinated.
Since pioneering the field of urban simulation and publishing “UrbanSim: Modeling urban development for land use, transportation, and environmental planning” in the Journal of the American Planning Association twenty years ago, a lot has changed about urban models and their usefulness. Years of research went into creating UrbanSim, funded by the National Science Foundation, MacArthur Foundation, EPA, and others. This phase of R&D led to the refinement and battle-testing of the modeling methods embedded in UrbanSim, through rigorous peer-reviewed publications and competitive grant applications. While having thousands of citations to publications of the UrbanSim modeling approach provided validation of the methodology, it wasn’t enough to demonstrate that the models could be useful outside the ivory tower.
When we began to take UrbanSim out of the ivory tower and into the real world for more practical tests of its usefulness, the process of building models was initially slow and painful, and the models were difficult to use, requiring a rare combination of programming, data management, and econometrics expertise. UrbanSim gained a reputation in those years for being a gold standard in terms of scientific validity but was criticized as being too data-hungry and difficult to build and use to be useful for smaller organizations. We listened to the feedback and began leveraging improvements in automation, software engineering, and user interface design to make the process of building models more efficient and to make the user experience more intuitive and productive. This led to the creation of a SaaS platform to enable cloud deployment of models and their operation through easy-to-use web browser interfaces. More on this in a later post.
The original questions that motivated the development of UrbanSim models remain just as motivating today. These are questions about where and when and how urban areas will develop, how those dynamics will shape demand for infrastructure and services, and in turn how the provision of infrastructure and variation in local development regulations will impact urban real estate demand, supply, and prices.
Those foundational questions have been augmented over time by closely related questions such as how infrastructure and development regulations can be leveraged to reduce greenhouse gas emissions and energy consumption, how transit investments and development policies can be more effectively coordinated to maximize their effectiveness, and how impactful alternative interventions could be to reduce the devastating impact of the pervasive and worsening housing affordability crisis.
2022 was my first full year to focus exclusively on making the UrbanSim models and platforms more broadly useful to public and private sector users, after closing my last chapter as an active faculty member and former chair of the Department of City and Regional Planning at UC Berkeley. As we come to the end of 2022, it is a good time to reflect on the questions that motivated the development of urban simulation as a field and to review how federal, state, metropolitan, and municipal governmental agencies are using UrbanSim models today to answer such questions. In 2022, three federal agencies in the U.S., two federal agencies in Canada, three U.S. National Labs, one U.S. State, two international consulting firms, two large municipalities, and fourteen U.S. Metropolitan Planning Organizations including most of the largest metropolitan areas in North America and Australia used UrbanSim models to address their challenges. Models have in fact come a long way from the ivory tower to making an impact in practical application to hard problems, at every level of government from the local municipality to the federal agency.
In subsequent articles and posts, we will provide a more detailed recap of 2022 at UrbanSim and explore in more depth questions about model usefulness and what we are doing to make urban models and data more broadly useful. Until then, best wishes and happy holidays!
