CMStatistics 2019: Start Registration
View Submission - CFE
A0251
Title: Entropy as measure of spatial agglomeration: Interactions of business locations and housing transactions Authors:  Katarzyna Kopczewska - University of Warsaw (Poland) [presenting]
Abstract: Entropy is usually used in measuring the concentration of business (often called specialisation) and applied to data aggregated by sectors and regions. This traditional approach does not respond spatial analyses challenges and should be extended for point data to measure the spatial agglomeration over space. This however requires transformation of data to obtain the weights of points. Voronoi tessellation tiles can approximate well the point pattern in the continuous space. The switch from point data to polygonal representation also changes the approach to spatial relation, mainly distance between objects. Instead of measuring the spatial separation between points, one can use the share of tiles surface in whole area to conclude about the closest neighbourhoods. This also opens the opportunities of using entropy, because of proportional character of percentage areal data. Thus, the tessellated point pattern can be examined for the existence of agglomeration with entropy measure. Comparative analyses on the location patterns and density of points are of particular interest in urban studies. However, aggregated data for urban areas erase many of spatial information, what lowers the analysis power in this very highly diversified environment. Urban studies benefit from low-granulation data, especially point data. Point data applied to entropy via tessellation can be used to understand how the spatial allocation interacts and to detect spatial agglomeration of point data.