Integrating machine learning and spatial analysis for business establishment characterisation

This study examines the spatial and operational characteristics of businesses in the Halifax Regional Municipality, Canada. A two-stage framework is used to link business characterization with spatial analysis. K-Prototype clustering groups establishments by sector type, employment, population density, and zonal entropy. Kernel Density Estimation and Ripley’s K-function are then applied within each cohort to examine spatial intensity and concentration. Three sector-structured cohorts are identified. Production, wholesale, and transportation activities are associated with lower-density contexts. Service activities are concentrated in high-density urban areas, while retail and other services are more common in suburban contexts. The spatial analyses further show differences in establishment intensity and regional concentration across the cohorts. These patterns help describe the spatial organization of economic activities that generate or attract passenger and freight movement and provide a basis for future analysis of transport demand, accessibility, mobility, and freight.