Complex Network Variability Analysis with Impact of Major Events on Aviation Networks

Global aviation data has complex, large and fast-updated characteristics which it's not suitable for directly building complex networks. It's challenging to assess the stability of a particular airport by analyzing the aviation network. This paper presents Complex Network Variability Analysis (CNVA) to analyze the stability of the nodes in the aviation network, which face to international incidents such as COVID-19. Firstly, CNVA cleans the data with missing information to solve the problem of data redundancy. We compress flights' data which have same origin and destination and count their number. Then, CNVA calculates all nodes' harmonic closeness centrality, betweenness centrality and eigen centrality of the data in each month, which can analyze the significance features of each months' network. Next, we calculate the centrality's average rate of variety, and we use these data's average of the absolute values as parameter value. The parameters are used to construct a vector function which called V function. HCC, BC and EC are the variables of V function. Finally, we evaluate whether a node has a large change during COVID-19 period, we calculate the difference between the average change rate of the node's V value and the average change rate of the global function V, and we evaluate its size. To verify the credibility of CNVA, we search for nodes with a higher degree of centrality for CNVA method verification. The results show that the node's V constant value greater than V function constant which have better fluctuations when they face to COVID-19. Our analysis illustrates that CNVA can better assess the stability of airports in response to major incidents when the raw data sources are sufficient.