Use of Open street map to population estimation for small areas in urban center
ISBN 978-85-88783-11-9
Authors
1França, V.; 2Strauch, J.; 3Ajara, C.
1IBGE Email: vitor.franca@ibge.gov.br
2IBGE Email: julia.strauch@ibge.gov.br
3IBGE Email: cesar.ajara@ibge.gov.br
Abstract
As GIS and remote sensing data availability, the principles of Semenov-Tian-Shansky (1928) dasymetric method could be applied to the problem of zonal interpolation digital data. A dasymetric map has more homogeneous areas derived from the intersection of origin areas and assistants that allows the reader to get a better understanding of the distribution of the variable when compared to choropleth maps. The interpolation dasymetric has several approaches and can be performed with different auxiliary information sources. Among these approach stands out the zonal interpolation method developed by Bufallino and Reibel (2005) which uses bases of roads and streets to transfer values between different zonings. This methodology is developed in two steps, where each street segment receives a weight according to the ratio between their length and the sum of the length of all segments contained in the source zone. The first step of this methodology the overlay three layers of vector maps: origin zones, destination zones and street axes. All axes are targeted at the edge of intersection between the origin and destination zones to prevent a feature are counted more than once. The second step assign a weight to each street segment. This weight is calculated as the ratio between its length and the sum of the length of all segments in the source zone sector. The population in the segment can be estimated by multiplying the weight of the original population in the sub-district, and in the target area by the sum of the estimated population in the segments contained therein. In this paper, this methodology is applied to the urban areas of the cities of Belém and Ananindeua. For this, are used data the street axes of the OpenStreetMap and population counts of 2010 Census aggregated by subdistrict. So it is possible to calculate the weights and estimate the population in each street segment. Then, the estimated population was aggregated by census tracts and compared with the known value. This method uses only vector data and it showed computationally simpler and easier to use without GIS knowledge. We also emphasize that this approach proved most effective in urban areas than in rural areas, where population density has no correlation with the necessarily high road density.
Keywords
dasymetric map; population estimation; urban area