Beyond Demographics: Introducing the Behavioral Census for Urban Modeling

Traditional census data gives us a snapshot of where people live, but it has its limits. In this talk, Hamish Gibbs from Northeastern University explores the concept of a "behavioral census", a new way to understand cities by focusing on what people do outside their homes. Discover how a small set of understandable behavioral features can represent how people access amenities and interact with the urban environment. This approach provides a richer layer of information that can complement traditional demographics and improve how we model and plan our cities for the future. Inspired by this clip? Watch the full talk and all sessions from SDSC NYC 2025 on demand: https://spatial-data-science-conference.com/2025/newyork#Watch-on-demand #SDSC25 #geospatial #spatialdatascience #urbanplanning Transcript: 00:03 I'd like to start today by talking about a census. We can start with this map. This is a map of median income in Boston from the US census at the census block group level. I think everyone in this room should be quite familiar with using census data. When we want to understand how people access amenities and opportunities in cities, our first port of call is data from the census. 00:25 This tracks lots of different population characteristics that then get included in lots of downstream modeling. The census is a valuable source of information, but it has some limitations. One of those is it's a complicated, expensive survey to conduct, so we get a census released every 10 years. 00:43 In the US, we get modeled updates annually through the American Community Survey. It is also really focused on the characteristics of residential areas—what are the characteristics of the people that live in these areas? We think that the focus on residential characteristics is a reason for many assumptions around a person's residence being important for what amenities they can access or their demographics being useful for explaining their behaviors. 01:09 The research I'm talking about today is focused on building a set of behavioral representations that can capture what people are doing outside of the places they live, but at a granularity similar to the census. We could use this extra information about how people are accessing amenities in cities to improve a lot of urban modeling. 01:32 The idea we're driving at is what we're calling a behavioral census. The idea here is that we have a small number of very interpretable behavioral features that we can use in a way that is similar to existing demographic features. Something like income, ethnicity, or education is very salient to policymakers or decision-makers. 01:51 In a similar way, we're interested in whether we can represent people's behavior outside of the home in a small number of understandable behavioral features.