Intercensal Updating Using Structure-Preserving Methods and Satellite Imagery

Abstract

Censuses are fundamental building blocks of most modern‐day societies, yet collected every 10 years at best. We propose an extension of the widely popular census updating technique structure‐preserving estimation by incorporating auxiliary information in order to take ongoing subnational population shifts into account. We apply our method by incorporating satellite imagery as additional source to derive annual small‐area updates of multidimensional poverty indicators from 2013 to 2020 for a population at risk: female‐headed households in Senegal. We evaluate the performance of our proposal using data from two different census periods.

Small Area Estimates of Poverty Incidence in Costa Rica under a Structure Preserving Estimation (SPREE) Approach

“…In order to apply a fitting strategy via IPF, Koebe et al (2022) summarise some basic requirements that should be considered:…”

“…The use of the EBP in each year of study with a final benchmark operation performed with MSPREE, 2. exploring potential advantages of using panel survey data or time-series models for example, with the extension of the Fay-Herriot model proposed by Rao and Yu (1994), or 3. implement measurement error models also in the context of arealevel models (Ybarra and Lohr 2008). Finally, it is recommended to study the inclusion of non-traditional information sources (e.g., big data) as proposed in Koebe et al (2022) since the structures of population censuses can quickly become obsolete. A clear example of this is the socio-economic effect that the COVID-19 pandemic generated in many countries, altering the living conditions of many people in a short period of time.…”

Small Area Estimation of Poverty Under Structural Change

“…Cluster‐averaged variables might also stem from different sources apart from census data. There is a growing literature showing that big data from mobile phones or remote sensing are readily available in many countries and predictive for up‐to‐date poverty measures at a high spatial resolution (e.g., Engstrom and Soundararajan, 2020; Koebe et al ., 2021; Jean et al ., 2016; Njuguna and McSharry, 2017; Pokhriyal and Jacques, 2017; Steele et al ., 2017). As will be shown, the inclusion of such data in the methods discussed in this paper is straightforward.…”

Experimentelle georeferenzierte Bevölkerungszahl auf Basis der Bevölkerungsfortschreibung und Mobilfunkdaten

“…von der Gebäudedichte, Gebäudehöhe und den Gebäudetypen aus Satellitendaten (Copernicus Sentinel-1 und Sentinel-2 Daten) hergeleitet werden. Neuere Ansätze wie in Koebe et al (2022) kombinieren Satellitendaten und das Small Area-Schätzverfahren SPREE 6 , eine Methode der strukturerhaltenden Schätzung, die insbesondere für regional und demografisch differenzierte Bevölkerungsfortschreibungen zwischen den Zensen auf kleinräumiger Ebene verwendet wird. Anhand der Hilfsinformationen aus den Satellitendaten werden folglich kleinräumige Bevölkerungszahlen für den Senegal selbstständig fortgeschrieben.…”