This lecture covers the determination of the spatial neighborhood of geographic objects using various criteria, the choice between fixed or variable kernels, the selection of contiguity order for polygons, and the importance of defining spatial weighting schemes. Understanding spatial autocorrelation involves comparing the behavior of a specific area with that of each geographic object, achieved through calculating the correlation between the variable distribution of each object and the average distribution within its neighborhood. The lecture also introduces the Moran's I index and explains the significance estimation through random permutations using the Monte Carlo method.