2026-04-17T08:32:53Z https://datacatalogue.cessda.eu/oai-pmh/v0/oai

1ccf0dc109ce3de5fe53d8295cb3c6731508d1e26f2f080aadd43b350f46dac1 2025-06-17T03:14:54Z language:enopenaire_data

DDI study level documentation for study 10.7802/1918 Human Lights version 1 Human Lights version 1 Human Lights version 1 10.7802/191810.7802/191810.7802/191810.7802/1918 Määttä, Ilari Määttä, Ilari GESIS Data Archive for the Social SciencesGESIS Datenarchiv für Sozialwissenschaften Satellite nighttime lights open new opportunities for economic research. The data is objective and suitable to study regions at various territorial levels. Given missing reliable official data, nightlights often proxy for economic activity in particular in developing countries. However, the commonly used product, Stable Lights, has problems to separate background noise from economic activity at lower levels of light intensity. The problem is rooted in the aim of separating transient light from stable lights, even though light from economic activity can also be transient. We propose a new method that aims to identify lights emitted by human beings. We train a machine learning algorithm to learn light patterns in- and outside built up areas using GHSL data. Based on predicted probabilities we include lights in those places with a high likelihood of being man-made. We show that using regional light characteristics in the process increases accuracy of predictions at the cost of introducing a mechanical spatial correlation. We create two alternative products as proxies of economic activity. Global Human Lights minimizes the bias from using regional information, while Local Human Lights maximizes accuracy. The latter shows, that we can improve detection of human generated light especially in Africa.Satellite nighttime lights open new opportunities for economic research. The data is objective and suitable to study regions at various territorial levels. Given missing reliable official data, nightlights often proxy for economic activity in particular in developing countries. However, the commonly used product, Stable Lights, has problems to separate background noise from economic activity at lower levels of light intensity. The problem is rooted in the aim of separating transient light from stable lights, even though light from economic activity can also be transient. We propose a new method that aims to identify lights emitted by human beings. We train a machine learning algorithm to learn light patterns in- and outside built up areas using GHSL data. Based on predicted probabilities we include lights in those places with a high likelihood of being man-made. We show that using regional light characteristics in the process increases accuracy of predictions at the cost of introducing a mechanical spatial correlation. We create two alternative products as proxies of economic activity. Global Human Lights minimizes the bias from using regional information, while Local Human Lights maximizes accuracy. The latter shows, that we can improve detection of human generated light especially in Africa. OtherOtherAndereOther Free access (without registration) - The research data can be downloaded directly by anyone without further limitations. CC BY 4.0Freier Zugang (ohne Registrierung) - Die Forschungsdaten können von jedem direkt heruntergeladen werden. CC BY 4.0 http://dbkapps.gesis.org/dbkoai/oai.asp oai:dbk.gesis.org:SDN/10.7802_1918 2023-03-11 ddi:codebook:2_5