Urban built environment, lifestyles, dietary habits, physical indicator and health: a machine learning analysis in Japanese cities

This repository provides the spatial datasets used in the study “Urban built environment, lifestyles, dietary habits, physical indicators and health: a machine learning analysis in Japanese cities”. The datasets were compiled and processed by the authors to support the empirical analyses examining the associations between urban built environment characteristics, lifestyle and dietary patterns, physical indicators, and health outcomes across multiple Japanese cities using machine learning approaches. The shared data primarily include spatially aggregated built environment indicators derived from open urban data sources (https://nlftp.mlit.go.jp/ksj/). These indicators capture multiple dimensions of the urban environment, such as land-use structure, density and accessibility of facilities, transportation-related characteristics, and environmental context. All variables were processed at consistent spatial units to enable integration with individual- or area-level health, lifestyle, and dietary information used in the analysis. To protect privacy and comply with data governance requirements, no individual-level or personally identifiable information is included in this repository. Health-related, lifestyle, and dietary variables are not shared at the raw microdata level; instead, this repository focuses on the spatial data and derived indicators that can be openly distributed and reused. The data provided here allow replication of the built environment feature construction process and facilitate reuse in related studies on urban health, spatial epidemiology, and machine learning–based urban analytics. Researchers may combine these spatial indicators with their own health or survey data to explore similar research questions in other contexts. This dataset is shared for academic research purposes. Users are requested to cite this dataset appropriately when using it in publications or derivative works.