Module 3.1 Scale Effect and Spatial Data Aggregation

This last lab discusses about scale effect and spatial data aggregation in both vector and raster data. For the scale effect in vectors, the relationship between different scales (1:1200, 1:24000, and 1:100000) is that length, perimeters and area decreases as you decrease the scale and vice versa. This might mean the level of detail decreases as the length and area decreases. In the readings (Goodchild 2011, p.6), as the resolution or scale becomes finer (increase) then calculated areas, volumes or lengths would increase as their boundaries are given more detail. Lines and polygons when you increase scale (1:1200) are finer than a lesser scale where they look generalized. Which makes sense because whenever you delineate polygons and lines using an imagery you need to increase scale or zoom in in order to have finer delineation compared to a generalized line. The generalization is good for faster render of data while the finer details are good for accurate analysis.

While rasters, cell size increase the level of detail within DEM (and upon converting to slope) decreases. In the DEM, the 90m resolution was pixelated compared to the 1m resolution. This makes sense because cell size refers to the distance of each pixel/cell to the next one and the value that is assigned per pixel/cell is defined per that resolution. As per Kienzle (2004, p86), slope values increase with a decrease in grid cell size – which is applicable to DEM as well.

We also discussed and test Modified Area Unit Problem using regression and saw that the impact of results of the OLS affects when point-based measure of spatial data are aggregated into districts areas. The summary values are influenced by the scale of the aggregation unit. 

Lastly, we did Gerrymandering which is an act of manipulating political boundaries in all forms of government and can be measured in severity by community and compactness of the physical boundary. There are 2 ways in determining the compactness and community of a district. For community, we know that according to multipart polygons – we want to minimize dividing it into multiple parts. Like above, we have 9 districts that have multiple separate polygons. Next, compactness is determined through the Polsby-Popper score wherein if the score is 1 then the more compact it is therefore farther than 1 we consider it as the “offender” in this case.

Step 1: Create 3 fields named Perimeter, Area and Polsby-Popper score

Step 2: Calculate geometry attributes for perimeter and area

Step 3: Calculate field for Polsby Popper score using the formula (12.56637 * !Area!) / (!Perimeter! * !Perimeter!)

Step 4: Sort is ascending to know who has the lowest score and we determine that the offender is: Congressional District 12 with 0.03 Polsby-Popper score. See the screenshot below: