Recall our definitions of the ecological fallacy;  modifiable areal unit problem.

Recall the questions I left the class with from Sherwood's chapter:
1. Be able to respond (in class, and on paper) to Sherwood’s generalizations about the state of geographic education and linkages to business communities, from your own experiences and from an institutional perspective
2. How do you [plan to] describe yourself professionally?
3. What does she mean by the phrase “business geographics”? [She gave no explicit definition, but I’d say “concepts and systems that link tools for manipulation of geographic information, in the service of business or commercial decision making”]
4. How does this differ from “economic uses of geographic information”?  From “GIS applications”?

What sorts of geographic analyses have you undertaken?
 

Birkin et al. 1996
GIS and Spatial Decision Support Systems
 

What GIS functions do the authors cite as being used in typical analyses?

• Overlay (what is it;  why would it be used?)
• Weighted overlay (what is it;  why would it be used?)
• Buffering (what is it;  why would it be used?)
• Point-in polygoning (what is it;  why would it be used?)
• Routing (what is it;  why would it be used?)
• Data transformation (arithmetic as well as geographic)  (what are these;  why would each be used?)

Spatial analysis:  using geographical information –

• to uncover spatial regularities (inductively generating theory or explanation),
• to test expectations (inductively testing theory), or
• to forecast for geographic areas among which there may be some interrelationship (e.g., forecasting migration flows, which reflect economic and social circumstances within regions and relative conditions between regions, with a distance-decay function;  or forecasting store sales based on local market characteristics by distance from the store, and competitors’ actions and market distance from competitors);  the forecast is based on a model which can be a priori or could have been estimated based on earlier data (explain);
• using geographic modeling to present or propose geographical explanations (presenting deductive theory) – geographic modeling of causal processes.
• Spatial autocorrelation (to uncover social processes – juvenile delinquency;  violent crime – to uncover economic correlations)

Thomas, 2000
Insurance Pricing with GIS
Modeling geographic patterns of risk

What kinds of risk should face geographic correlation, at what geographic scale for each?

• Fire damage – at the localized scale (this is why fire protection is a public good rather than privately provided)
• Earthquake damage – at the global scale of plate tectonic, and the more localized scale of soil types
• Hurricane damage – at the broad regional scale of hurricane incidence and pathways (hazard), and the more localized scale of topography and municipal building codes (vulnerability) – though what Thomas suggested as a proxy for vulnerability was the company’s proprietary data on loss claims by small area.