Hi all,
I am working on a project to crosswalk data from the census tract level to the state legislative district level. I am doing this by calculating the geographic overlap between tracts and state districts and using tract-level census data (datasets like total population, population by race/age, etc.) to construct an average at the state district level.
I am running into an issue where tract-level census datasets appears to have more tracts than the GIS files at the tract level. For example, the year 2000 Total Population dataset at the tract level has more tracts than the year 2000 tract level GIS files.
Currently, I am working around this problem by just eliminating the missing tracts from my calculations, but I was wondering if anyone knows the cause/has any other solutions for this problem. Thanks!
We clip NHGIS shapefiles at the coasts and Great Lakes shores, which deletes some tracts that are entirely offshore. There is no population in any of these deleted tracts, so their absence shouldn’t impair a population crosswalk.
NHGIS also doesn’t include Puerto Rico in its 2000 shapefiles or tables. If you’re getting census tables from another source, they may include Puerto Rico.
Besides that, we highly recommend you consider using data for units smaller than tracts, such as blocks, to crosswalk data to other areas. Where tracts are split by state legislative district boundaries, your tract-based allocation will cause errors that may be unnecessary if you start from smaller units. See this explanation on the NHGIS crosswalks page, which also supplies some crosswalks that may help you with your application.