ABSTRACT The spatial distribution of economic activities is central to regional economics. However, empirical tools for comparing industries based on the similarity of their spatial distributions remain limited. Traditional cluster analysis typically ignores geographical distance and connectivity, while spatial methods focus on clustering places rather than economic activities. This paper addresses this methodological gap by proposing a framework for clustering economic activities based on the similarity of their spatial distributions, explicitly incorporating geographical space. The Wasserstein distance is employed with a geographically informed cost matrix, that is, the Shimbel matrix, to measure the effort required to transform one activity's spatial distribution into another. This approach is compared with traditional non‐spatial distance metrics through an illustrative example and then applied to 272 agribusiness activities across 466 municipalities in Brazil's Midwest region from 2013 to 2021. The results reveal four main clusters: one spread throughout the region and others concentrated in northern, southern, or eastern areas, consistent with regional endowments, historical trajectories, and agro‐industrial linkages. The case study suggests that the approach offers a useful complementary tool for identifying co‐distributed activities and spatially similar industrial patterns, with implications for regional development analysis, place‐based policy design and future studies of economic linkages and shock exposure.
Júnior et al. (Wed,) studied this question.