Optimizing the spatial point sampling for population migration dynamics modeling in Siargao Island, Philippines

Abstract

Simulating population migration dynamics on a geospatial network requires the network to match street design standards applied to a specific region. Using OpenStreet Map data of mainland Siargao, a network of possible population settlement points is obtained using different distance resolution θ between points avoiding realization of fully connected subgraphs. To determine an optimized sampling, the tolerance θ value of the projected graph was varied in discrete values from 1 to 250 m where the node count n, edge count m, and average clustering coefficient C were calculated and plotted against θ. We identified θ = 34 m to be the optimal tolerance value since it is the closest value to the minimum positions of the best fit curves in the delta plots of n and m and lies within the great variations of the plot of θ vs C. The resulting street network is comprised of 281 nodes and 914 edges which will be used in the geospatial simulation of population migration dynamics.