Table 1.
Zip4 record with attributes.
Table 2.
An example of voter data with real, Zip4, Street Segment, Census Block Group, and Zip centroid details.
Fig 1.
Spatial distribution of length of Zip4 segments for (a) Cuyahoga County, and (b) Portage County.
Zip Code boundaries are also displayed.
Fig 2.
Kernel Density Estimate (KDE) surfaces generated for (a) Real, (b) Zip4, (c) Street Segment, (d) Zip, and (e) Census Block Group.
Fig 3.
SaTScan clusters for (a) Real, (b) Zip4, (c) Street Segment, (d) Census Block Group, and (e) Zip.
Fig 4.
GeoMEDD clusters for (a) Real, (b) Zip4, (c) Street Segment, (d) Census Block Group, and (e) Zip.
Fig 5.
Spatial K-anonymity calculations.
The red point is the original location and the blue point is the geomasked location. The green stars are other potential locations.
Table 3.
Zip4 length distribution (in meters) for Cuyahoga and Portage counties.
Table 4.
Distance distribution (in meters) between real location and Zip4 centroid for Cuyahoga and Portage counties.
Table 5.
Average Nearest Neighbour analysis results for real and centroid datasets.
The NNR is calculated as the average for 100 datasets.
Table 6.
Ripley’s K results using the real and centroid datasets.
Table 7.
KDE raster cell comparison between real and centroid datasets.
Table 8.
Similarity analysis results for SaTScan and GeoMEDD clustering.
Fig 6.
K value distribution with respect to the percentage of total addresses for (a) Zip4, (b) Street Segment, and (c) Census Block Group.