Someone raised a good point today about the frequency of changes for the boundary lines of state legislative districts. A primary concern is whether they change less frequently (and how much so) vs the updates made to the TIGER shapefiles provided by the Census Bureau. We could risk some really stale data if the changesets are large and frequent.
One way to mitigate this is to use the voter file itself as a way of determining where the boundaries lie, basically reverse engineering the voter file into regions
Apparently even mapping zip-9s to political geography could have something like only an ~2% error rate. The likelihood of a zip-9 crossing relevant political boundaries is supposedly relatively low.
Someone raised a good point today about the frequency of changes for the boundary lines of state legislative districts. A primary concern is whether they change less frequently (and how much so) vs the updates made to the TIGER shapefiles provided by the Census Bureau. We could risk some really stale data if the changesets are large and frequent.
One way to mitigate this is to use the voter file itself as a way of determining where the boundaries lie, basically reverse engineering the voter file into regions
Apparently even mapping zip-9s to political geography could have something like only an ~2% error rate. The likelihood of a zip-9 crossing relevant political boundaries is supposedly relatively low.
We should also really figure out how much headway has been made at the Voting Information Project.