Lifestyle embeddings reveal multi-scale behavioral change in US cities | Northeastern University

In this session from the Spatial Data Science Conference New York 2025 (#SDSC25), Hamish Gibbs (Postdoctoral Researcher, Social Urban Networks Lab at Northeastern University) presents research on using large-scale mobility data to build a “behavioral census," a framework for capturing how people interact with cities beyond where they live. He explains that traditional census data focuses on residential demographics, while mobility data reveals real-world behaviors such as commuting, shopping, and dining. Using anonymized GPS data from 20 U.S. cities, he identifies “lifestyles” as shared patterns of daily activity that help explain differences in urban outcomes and how they’ve evolved through events such as the COVID-19 pandemic. 0:08 Introduction & overview of research focus 01:28 Limitations of traditional census data 03:09 The concept of a behavioral census 04:01 Identifying lifestyles from mobility data 07:13 Key findings 10:20 Tracking lifestyle changes during and after the pandemic