GIS Workshops - Spatial Regression



  Intro to GIS  |   Coordinates Systems  |   Exploratory Spatial Data Analysis  |   Points and Surfaces  |   GWR  |   Spatial Regression  |   Spatial R

In regression analysis we usually assume our observations are independent of one another. However, for many processes we might expect to see interaction among observations that are spatially proximate to one another: people live in one county but cross into another to work or shop, increased police activity in one neighborhood may increase (or decrease) criminal incidents in adjacent places depending on conditions. The presence of spatial dependence in our data, as these kinds of relationships are collectively known, violates key assumptions of regression analysis. Spatial Regression encompasses several techniques available to incorporate spatial dependence into our models including spatially lagged dependent and independent variables and spatially lagged error terms. Using the open source software package GeoDa we will diagnose spatial dependence in a data set on county-level poverty and develop a set of spatial models that leverage this relationship to improve the reliability of our original model.

Prerequisite: The workshop assumes some prior knowledge of GIS. In particular, the workshops on Exploratory Spatial Data Analysis and Geographically Weighted Regression are highly recommended.

Instructor

Chris Fowler
Raitt Hall 218M
csfowler@uw.edu
(206) 920-1686

Outline

  • What is Spatial Regression and why
    should we consider it?
  • Spatial heterogeneity and spatial dependence
  • Spatial autocorrelation
  • Spatial Error models
  • Spatial Lag models
  • Interpreting Spatial Regression models

Materials

Additonal Resources

  • Anselin, L. 2002
    • A primer on the spatial processes whose presence can motivate spatial regression. Also covers weights matrices and their tradeoffs.
  • Anselin, L. 2005
    • A primer on spatial regression using R.
  • Elhorst,J.P. 2010
    • An economist's view on the state of the art in spatial regression. Includes a useful taxonomy of spatial model forms and their uses.
  • Voss P.R. et al. 2006
    • An exceptionally clear empirical application of spatial regression from the discipline of Sociology. Very clear about the modeling choices throughout the analysis.

To register for the workshops, please complete the Registration Form.

Spatial autocorrelation in the residuals of a standard OLS Regression (Poverty in the Southern U.S.)



Examining significance of spatial autocorrelation
in GeoDa