Using light-at-night (LAN) satellite data for identifying clusters of economic activities in Europe

Abstract

Enterprises organized in clusters are often efficient in stimulating urban development, productivity and profit outflows. Identifying the clusters of economic activities thus becomes an important step in devising regional development policies, aimed at the formation of clusters of economic activities in geographic areas in which this objective is desirable. However, a major problem with the identification of such clusters stems from limited reporting by individual countries and administrative entities on the regional distribution of specific economic activities, especially for small regional subdivisions. In this study, we test a possibility that missing data on geographic concentrations of economic activities in the European NUTS3 regions can be reconstructed using light-at-night satellite measurements, and that such reconstructed data can then be used for cluster identification. The matter is that light-at-night, captured by satellite sensors, is characterized by different intensity, depending on its source—production facilities, services, etc. As a result, light-at-night can become a marker of different types of economic activities, a hypothesis that the present study confirms. In particular, as the present analysis indicates, average light-at-night intensities emitted from NUTS3 regions help to explain up to 94 % variance in the areal density of several types of economic activities, performing especially well for professional, scientific and technical services $$({R}^{2}=0.742{-}0.939)$$ , public administration $$({R}^{2}=0.642{-}0.934)$$ , as well as for arts, entertainment and recreation $$(\hbox {R}^{2}=0.718{-}0.934)$$ . As a result, clusters of these economic activities can be identified using light-at-night data, thus helping to supplement missing information and assist regional analysis.