IDEM 356

Spatial Demography

Concepts, Spatial Statistics, GIS and Cartographic Techniques

Start: 5 November 2013
End:  31 January 2014
Location:  MPI Rostock

Dr. Sebastian Klüsener

Course description:
Researchers often work with geographically referenced data. In comparative social demographic research it is very common to contrast populations across countries or regions. Biodemographers, on the other hand, often have information on the geographic habitat of animals or plants they study. One can represent this information in non-spatial tabular form and analyze it with standard statistical techniques that do not make use of the spatial information. However, by ignoring spatial information contained in the data, spatial relationships, spatial trends and spatial contextual effects remain unexplored and underutilized. Analyzing spatial data, geographical maps can help to get a first understanding of the data. But patterns in geographical maps may often not be significant because they are simply the outcome of the intrinsic variability of a phenomenon. Significance tests are hence crucial before jumping to conclusions. Modeling geographical data, the exclusion of spatial information can even lead to biases in the statistical models where some of the basic modeling assumptions may be violated. Therefore, understanding the spatial processes underlying the relationships of interest can improve overall knowledge of demographic events as well as improve the usefulness and applicability of statistical models.

The course will give an introduction to techniques and programs used in the field of Spatial Demography. It covers methods that are useful for social demographers as well as techniques that can be applied in biodemography. The course will start with an overview of concepts and theories used in Spatial Demography. This is followed by a brief introduction to Geographical Information Systems (GIS), spatial data files and the spatial libraries in the software package R. Course participants will then be getting an overview over tools of descriptive analysis and cartographic presentation as well as basic and advanced spatial modeling techniques. Thereby, methods to analyze vector data (e.g. countries, regions), point data (e.g. count data of human individuals, centroids of regions) and raster data (e.g. satellite images of land use) are covered. This includes Spatial Lag/Spatial Error Models, Spatial Multi-Level Models and Spatial Poisson Regression Models.

This course will be a practical course composed of one lecture, ten lab sessions and two classes where the participants present findings of mini research projects, which they carried out during the course. The course will be offered between the 5th of November 2013 and the 31st of January 2014. We will use the following software packages: R, Geoda, Quantum GIS. Attendance is limited to a maximum of 15 people. Students should plan to spend about 4-5 hours of work-time per week on this course.

Course prerequisites:
Participants should be familiar with basic multivariate analysis techniques (linear and logistic regression, test of significance, confidence intervals). Prior knowledge of Geographic Information Systems, spatial statistics and cartographic techniques is not required. Some basic knowledge of R is desirable, but not a prerequisite.

Students are expected to submit a mini-project upon completion of the course. For the project students can either analyze data related to their own research OR they can submit a project based on datasets that are provided by the course instructor.

Course material:
Lecture notes, material and datasets will be made available during the course.

General readings:
Anselin, L. (1988): Spatial Econometrics: Methods and Models. Dordrecht.
Anselin, L. (2005): Exploring Spatial Data with GeoDa : A Workbook. Urbana-Champaign.
Cressie, N. and C.K. Wikle (2011): Statistics for Spatio-Temporal Data. New York.
Tobler, W. (1970) A computer movie simulating urban growth in the Detroit region. Economic Geography, 46(2): 234-240.
Voss, P.V. (2007): Demography as a Spatial Social Science. In: Population Research and Policy Review 26: 457-476.

How to apply

For application instructions please visit the applications page