This course explores spatial econometric models, which sit at the intersection of spatial analysis and econometrics. Spatial econometric models are appropriate for data where the units are not independent, but instead exhibit some type of spatial autocorrelation. These types of models are particularly useful for geographic data where the data from the geographic units are spatially dependent in some way. The effects may arise from externalities, interaction, or spatial concentration.

The course will begin with a discussion of why one might consider using these particular models and how they differ from other statistical approaches. Prior to this course, you should have already studied the linear regression model with some depth. We will examine what types of data are amenable to these spatial econometric approaches and what assumptions underlie the various methods.

The reading material will be dense. You will need to read carefully and slowly. The first reading will be a review piece that surveys the literature and provides a large picture of the field. Once we establish this background, we will then move on to applications of spatial econometric methods. I expect you to read the applications very carefully with close attention to detail, especially the methodological details. Carefully look over the table entries and the text describing the analysis presented. Please allot sufficient time to read each piece carefully. I cannot emphasize enough that none of the readings in this course should be skimmed.

We will end with a software session where I will demonstrate how to use various software packages for spatial analysis. In particular, we will explore GeoDA, which is a free software package that provides spatial data analysis tools, geovisualization tools, as well as a spate of tools for analyzing spatial autocorrelation and conducting spatial modeling. We will also examine qGIS, a free and open-source cross-platform desktop geographic information system (GIS) application that supports the viewing, editing, and analysis of geospatial data. ArcGIS is a similar tool, but is not free or open source. Your knowledge of qGIS will translate almost seamlessly to skill with ArcGIS. Lastly, we will examine several R packages that support the analysis of spatial data. We will cover the spdep package and may examine other packages as well, as time allows.