Spatial organisation and accessibility: a study of U.S. counties

The aim of spatial autocorrelation is to inspect the organisation and structure of spatial phenomena. This statistical method allows one to scrutinise spatial patterns and relationships that generate relevant phenomena in territories under consideration. In fact, spatial phenomena are often self-determining and may positively or negatively influence adjacent units. Autocorrelation analysis pinpoints the mutual influence among spatial units as well as any polarising effect around a specific area. In this manuscript we use spatial autocorrelation analysis to scrutinise the level of spatial dependency of accessibility for commuters in continental U.S. counties on the theoretical basis described above. Counties’ accessibility is computed according to gravity theory (spatial interaction models with impedance function obeying to exponential and power decay impendence functions). The scope of our study is to understand if space influences counties’ accessibility. We first apply global and local univariate spatial autocorrelation analysis in order to test the hypothesis that counties with higher accessibility positively influence adjacent counties. Finally we test whether accessibility is spatially correlated with socio-demographic variables, such as residential population and income per capita. We develop a bivariate autocorrelation analysis to assess spatial dependencies between U.S, counties’ accessibility and socio-economic variables both at the global and at the local level.