Why Geography is Important — Part II, 14 January 2015
When asking questions about geography and how to visualize it, it is important to think about spatial context. This brings up questions such as scale factors, MAUP, and appropriate data, among other things.
Different types of geography use different types of spatial units. The choice of unit can greatly affect the outcome of the analysis. The geographer must therefore be very thoughtful and careful when making choices. Human geography, for example, uses arbitrary units. Manipulation of these units can lead to issues with temporal analysis of census data or Gerrymandering. Natural physical units have their own issues, as they may have fuzzy or fractal borders.
As discussed in the last post, MAUP — the Modifiable Areal Unit Problem — is a common problem resulting from scale and aggregation issues. Once the issues have been identified, one must also how well the results can be generalized, or if they are specific to a certain study. The problem occurs because there are reasons that certain individuals are located where they are.
Neighborhood models are important ways of looking at grouped data. There are three types of models: grouping, group-dependent, and feedback. In the grouping model, individuals do not randomly move to certain areas, so they will form certain communities based on a desire to reside in specific locations. In the group-dependent model, individuals are subject to similar external influences. In the feedback model, individuals interact and influence one another, which will lead to the development of common characteristics. These models could be operating at different scales. Simpson’s Paradox may also be affecting spatial data in certain ways that become either clear or hidden as aspects such as scale are changed.