Nonlinearity is all you need! CHOAMs for better cities

Purpose This study advances a theoretical shift from linear to nonlinear modeling in urbanism, and introduces and empirically validates a nonlinear, weighted diversity framework, referred to as chains of activity models (CHOAMs), which links mobility, diversity and value. Design/methodology/approach This study is both theoretical and empirical. It performs quantitative research based on urban data collection (>179,000 entries), isochrone computations (>100,000 isochrones covering a 20 km × 20 km area) and modeling through neural networks. It discusses the development of the CHOAMs framework. Findings Urban value can be represented as weighted diversity, and model weights can be trained with CHOAMs, which can guide investors in building diverse cities, avoiding gentrification and other urban ills. Originality/value Building on preliminary results shown at SASBE 2024, this study provides an expanded discussion of nonlinearity and CHOAMs.