This site uses different types of cookies, including analytics and functional cookies (its own and from other sites). To change your cookie settings or find out more, click here. If you continue browsing our website, you accept these cookies.
It sounds like you're on the right track. I don't have time to build out an example, but here area few things to think about:
- One way to ensure a single ZIP match for each grid would be to useaSpatial Info tool to assign a centroid to each grid. Then a Spatial Match could be used to match those grid centroids to ZIP polygons.
- An alternate matching method would be to find the ZIP code that captures the greatest percent of land area of each grid and assign that grid to that ZIP code. Use the Spatial Match toolto output intersection objects and measure land area of each using the Spatial Info tool.
- Many data providers have different polygons for ZIP codes. It appears that your data specifically identifies "ZCTA" or the ZIP Code Tabulation Areas produced by the Census Bureau. I would suggest using the Census Bureau's TIGER linepolygonsfound at the link below for spatial applications in case there are any discrepancies between their polygons and those provided by Experian.
- Some ZCTAs in your data specify "TX Portion" meaning these polygons cross the state line. Consider using a polygon of the state of Texas to cut (Spatial Process tool) the polygons you use to only the Texas portion so your map only falls within Texas state lines.
- Depending on the Allocate data you have access to, consider using different base to distribute your employees. Rather than the entire population,maybe only the adult population of a working age is relevant here. Also, if you know the year this employee data wasmeasured, use the same data vintage of Allocate data.
Another way you could do this would be to use census block data for population weighting. The US Census Bureau publishes population and household counts at the Census Block level, in total there are about 11 million blocks across the US.
All you would need to do is generate the centroid for each census block, then intersect the block centroids, ZCTAs, and the grid. Find the total population of the ZCTA (looks like you already have that), as well as the proportion of the ZCTA population that resides in each grid cell (using the block centroids), then multiply your ZCTA labor data against the proportion of the ZCTA population in each grid cell. That should give you a fairly accurate way to weight data from ZCTAs to your grid.
This is exactly how the Allocate engine works in Alteryx, and you can build out the same functionality easily in a workflow.