5. Discussion and conclusion
In this study we have suggested a methodology for evaluating passengers’ transit service quality satisfaction and treating spatial variation of service quality attributes across the study area. Particularly, we have focused on transit service quality evaluated at railway stations. The research was based on the data collected by means of a survey conducted by the railway service operating in the hinterland of Milan. Users were asked to express importance and satisfaction rates about service quality factors revealed on board and at the station where the trip starts. Among them, seven attributes concerning the service quality at station were selected, elaborated and included in the analysis.
The main objective of the paper was analysing the spatial variation of transit service quality evaluated at station. For this aim, we have elaborated a spatial regression named geographically weighted regression (GWR) in order to quantify how the influence of each attribute on the satisfaction with the overall service quality varies across the study area.
Firstly, we have applied techniques of spatial statistics in order to establish if the spatial pattern of data is the result of random choice or of spatial association. Global and local measures of spatial autocorrelation were applied on the attributes of transit service quality and implemented in a GIS environment. The outcomes of these spatial statistics have demonstrated that the values distribution of the attributes of service quality present spatial autocorrelation. As a consequence, no-spatial models seem to be inadequate and spatial regression model, as GWR, could be applied.