GIS5935 M3.1: Scale & Spatial Data Aggregation
In this lab, we explored how scale affects spatial analysis by comparing regression results across different vector units: block groups, ZIP codes, counties, and House districts. Each unit produced different slope and R-squared values, showing that the relationship between race and poverty changes depending on the level of aggregation. This is a clear example of the Modifiable Areal Unit Problem (MAUP), where the choice of spatial unit can influence statistical (or personal) outcomes and interpretations. Smaller units like block groups captured more local variation, while larger units like counties tended to smooth out differences.
For raster data, resolution plays a similar role to vector data. Higher-resolution rasters provide more detailed information but can also introduce "noise" or complexity. Lower-resolution rasters simplify the data but may miss important local patterns. In both cases, choosing the right scale or resolution depends on the goals of the analysis.
Gerrymandering is the manipulation of political district boundaries to favor a particular party or group. It can be measured using compactness scores like the Polsby-Popper index, which compares a district’s area to the square of its perimeter. A score closer to 1 means the district is compact; a score closer to 0 means it has a more irregular shape.
In this lab, we calculated Polsby-Popper scores for each House district and identified the ones with the lowest values as the worst offenders. These districts deviate the most from what would be considered fair or unbiased boundaries. I included a screenshot of the "worst" district to show how far its shape strays from compactness.
Comments
Post a Comment