Munich Personal RePEc Archive

Researching commuting to work using the methods of complex network analysis

Pálóczi, Gábor (2016): Researching commuting to work using the methods of complex network analysis. Published in: Regional Statistics , Vol. 6, No. 1 (October 2016): pp. 3-22.

[img]
Preview
PDF
MPRA_paper_74496.pdf

Download (1MB) | Preview

Abstract

In the current paper the possible utilization of complex network analysis in spatial researches was investigated. The organizational and developmental regularities of networks were demonstrated from the aspect of regional development planning. The reviewed regularities provide a new approach of the regional developments. The dependencies of settlements were analysed with the application of disparity method on the basis of the commuting matrix of the census from 2011. The disparity of out-commuting exceeded the level of in-commuting in all population categories, producing a more significant dependency relation in case of out-commuting. In general, the value of disparity increases with decreasing population number in settlements and dependency grows. This can be related with decrease in the level of degree and commuting distance. According to detailed results, the method of disparity might be effectively used in additional spatial analyses as well. The community detection procedures of the complex network analysis were also applied for spatial division. Modularity optimization with the Louvain method was successfully used in the delimitation of larger territorial units. Smaller units can be created by the increase of the resolution but modularity stability deteriorates. At the same time the composition of the units changes. In the light of the results, it could be stated that regions formed by commuting relations (according to the process of regionalism) did not match the Hungarian NUTS2 statistical regions, but natural borders and NUTS-3 level administrative boundaries could be detected in more cases. The differences between the results and NUTS-3 boundaries are not unique distortions caused by the methodology but these reflect real commuting relations (the local labour system units were discussed in a previous study). The methodology might be appropriate to detect the hierarchical order of the local labour system’s units. The method is adaptable for additional analysis of spatial interactions.

UB_LMU-Logo
MPRA is a RePEc service hosted by
the Munich University Library in Germany.