Assessing the spatial and environmental characteristics of rural tourism lodging units using a geographical weighted regression model

Abstract

This paper presents a methodology to identify some factors influencing on the tourism market and not usually included in empirical analyses, such as those related to environment and location. The traditional quantitative analysis of spatially varying relationship assumes that the interdependence among variables measured at different locations is constant over the space. This assumption does not fit the data when the analysed variable presents spatial dependence. To face this problem, Geographical Weighted Regression (GWR) may be considered. The methodology proposed in this paper combines a genetic algorithm to automatically select the factors that best explain the dependent variable and GWR to determine the local estimations of the coefficient of regressors. A hedonic price model to analyse the rural tourism market in the island of La Palma (Canary Islands, Spain) was estimated in the study case. The results show that significant regressors are not homogeneously distributed throughout the island. Instead of a constant value, maps of values of the coefficients were obtained. These maps may be helpful to householders in order to implement local actions based on the attributes of the rental price of every house and estimate the economic returns of new rural houses sited in specific areas of the island.