This contribution is an introduction to the main topics of spatial econometrics. We start analyzing the main problems raised by spatial data. The first one is heterogeneity : statisticians must take account of the fact that spatial units may not be directly comparable. They must correct for differences in size, form, structure and so on. The second one is interaction among units located in space, the intensity of which decreases with distance. These interactions lead to spatial autoregression and spatial autocorrelation. Then, the paper introduces to the main instruments used to represent and analyze spatial autocorrelation and autoregression : spatial graphs, weight matrices, contiguity matrices. It presents the main tests used to detect spatial autocorrelation, color tests on qualitative data, Moran and Geary tests for quantitative data. It shows how these tests can be interpreted. An illustrative example is also provided. Last, the paper shows how to deal with spatial autocorrelation and autoregression on the example of linear models. The main types of spatial linear models are presented : spatially autoregressive, spatially autocorrelated and their combination. Then, we explain why least squares methods are not well suited to estimate this type of models. Most often, econometric analysis will rest upon maximum likelihood methods. The paper shows how to use these methods in the specific context of spatial models, in order to find parameters estimates and to make tests on them.