Abstract Knowledge-based economic activities (aka quaternary industries or QIs) are characterized by high concentrations of labour force and potentially high night-time light emissions. Therefore, geographic concentrations of such activities can presumably be identified using information on the amount artificial light at night (ALAN), which different geographic areas emit. Question, however, remains whether the models, incorporating ALAN data, are place-specific or whether such models are sufficiently generic, thus making it possible to apply them, once estimated, to other countries and continents. To answer this question, the analysis is performed in several phases. First, we build separate models for European NUTS3 regions and US counties. Next, we cross-validate these models and use them to predict QI concentrations worldwide. As the analysis shows, cross-validation of the models, applied to the “counterpart” continent, also results in a reasonably good fit, with R2 reaching 0.852, when the US-model is applied to the EU data, and R 2 = 0.896, when the EU-model is applied to the US data. Although attempts to use ALAN data for the analysis of different socio-economic phenomena are not new, to the best of our knowledge, this is the study first that uses cross-continent validation of ALAN-based models to determine their generality.
Testing the generality of economic activity models estimated by merging night-time satellite images with socioeconomic data
Advances in Space Research ; 66 , 11 ; 2610-2620
2020-09-02
11 pages
Article (Journal)
Electronic Resource
English
Real-Time Processing of Night Vision Images
British Library Online Contents | 2001
|Generality and legibility in mobile manipulation
British Library Online Contents | 2010
|Payment apparatus for electric vehicle charger with generality
European Patent Office | 2019
|