A Geographically Weighted Poisson Regression Approach for Analyzing the Effect of High-Speed Rail on Tourism in China

In the international literature, several studies have analyzed the impact of HSR on tourists' behavior with qualitative and quantitative approaches. However, they have not been able to solve the problem of capturing the spatial and temporal variation by fitting a regression model at a local point. The spatial heterogeneity within local models, such as Geographically Weighted Regression (GWR) models, provides a better platform allowing exploring the different spatial relationships between HSR and tourism. In this chapter, a spatio-temporal analysis has been proposed to evaluate the variables affecting tourists ‘choices, specifically the impact of HSR on both Chinese and Foreign tourists. Two advanced methods were adopted: firstly, we used the Weighted Regression with Poisson distribution (GWPR) modelling approach, which considers the problem of the temporal and spatial autocorrelation differently with respect to the Generalized Estimating Equations method. The results of this study support the use of the GWPR as a promising tool for tourism planning, especially because it makes it possible to model non-stationary spatially counting data. As far as the authors know, this methodology has never been applied in the international literature to this context. Secondly, we combined both temporal autocorrelation and spatial autocorrelation by applying models of Geographical and Temporal Weighted Regression (GTWR) types to take into account the local effects from the temporal point of view.