Evaluation of the Turkish Highway Network Analysis With Traffic Data

As a complex geospatial structure, Turkish national highway transportation network is studied by the means of network science. We used the dataset retrieved from the KGM (Karayollari Genel Mudurlugu) maps with a hand-driven process. The dataset labels the junctions in the map as nodes, and the roads between these junctions as edges. We outlined the statistical properties of the Turkish highway transportation network by the means of eigenvector, betweenness, closeness centrality, modularity and eccentricity measures, while comparative percentile plots between these measures are also performed. We investigated the correlation of these parameters with the traffic volume, and outlined that only eccentricity measure is correlated with the traffic volume. We also investigated the degree correlations of the network and found that the network displays disassortative mixing behavior, meaning that nodes with high degrees tend to connect with lower degree nodes, and vice versa. This property is consistent with the recent studies of transportation networks, as well as various types of real networks like Internet, World-Wide Web, protein interactions, neural network etc.