@zsarnoczay - suggests calculating a Mean Abs Error by block as a way to show quality of Building Inventory
Error based on the number of address points estimated by the Census Bureau versus the number in the building inventory.
This work is somewhat done in the file huest_v2-2-0_[focalplace]_2010_NSI.csv
This file is generated during the address point generation steps.
Suggestions for this file:
- add building sq footage
- add huestimate in each round
- summarize by block id
The additional step @npr99 would like to add is to look at socio demographic characteristics for blocks with larger uncertainty in the allocation.
This could also be done with the error check variables. Does the building inventory have less error at early stages of estimation correlation with income and tenure? Evidence suggests that building inventories tend to be more accurate in newer single family developments.
@zsarnoczay - suggests calculating a Mean Abs Error by block as a way to show quality of Building Inventory
Error based on the number of address points estimated by the Census Bureau versus the number in the building inventory.
This work is somewhat done in the file huest_v2-2-0_[focalplace]_2010_NSI.csv
This file is generated during the address point generation steps.
Suggestions for this file:
The additional step @npr99 would like to add is to look at socio demographic characteristics for blocks with larger uncertainty in the allocation.
This could also be done with the error check variables. Does the building inventory have less error at early stages of estimation correlation with income and tenure? Evidence suggests that building inventories tend to be more accurate in newer single family developments.