Automatic Generation of Generalized Population Density Maps from Raster Data
ISBN 978-85-88783-11-9
Authors
1Grötsch, A.; 2Schenkel, R.; 3Hurni, L.
1ETH ZÜRICH Email: baline@ethz.ch
2ETH ZÜRICH Email: schenkel@karto.baug.ethz.ch
3ETH ZÜRICH Email: lhurni@ethz.ch
Abstract
The analysis of the population dispersion is an often teached topic in geographical, historical or socio-economical education. The cartographic visualization of this subject is often implemented with the so called population density maps. They quantitatively show the number of people for a defined area – mostly administrative units – in an easily appreciable overview to the students. In those representations, however, it is assumed that the population is evenly distributed across the defined administrative unit. But in reality areas like parks, agricultural land and forests are sparsely populated or even uninhabited. Hence often a dasymetric approach is used to map population density, which combines census and land use data. For the 2016 edition of the SWISS WORLD ATLAS, the most used printed school atlas in Switzerland, all the existing small-scale population density maps need to be updated and harmonized. However not only the procurement of actual, worldwide census data is a challenge, but also the whole data processing from the source raster dataset to the final population density map with generalized vector data. In this paper the newly developed method to efficiently update the population density maps of different scales in the SWISS WORLD ATLAS will be shown. The developed algorithm is based on both, a global raster dataset from the Global Rural-Urban Mapping Project for the population, and the Corine Landcover dataset. Both of those datasets have a resolution of 3’ and are used in an expedient and sophisticated way during the process. The workflow adapts a clustering method and features an integrated generalization and vectorization for a reasonable map representation in the upcoming 2016 edition of the SWISS WORLD ATLAS. This paper includes further on recommendations for a best practice application of the algorithm as well as map examples of different scales.
Keywords
Density Map; Daymetric Map; School Atlas