twb

Creating regions of counties by work flow

Discussion created by twb on Jan 29, 2013
I suspect this may be a spatial analyst question. 

I have a polygon file of counties with three fields. 
Field A: Unique identifier of "home" county of where workers are coming from.
Field B: Unique identifier of "work" county of where workers are going to.  Of course, in many cases, the "home" county and "work" county are the same. 
Field C: Number of workers traveling from the "home" county to the "work" county.

This is a many-to-many table.  There's a separate row for County Y sending workers to County X, and another row for County X sending workers to County Y.  So, it's quite normal each county could be a listed in multiple rows as a "work" county, and in multiple rows as a home county. 

I want to have GIS cluster/regionalize the counties together based on their work flow relationship.  "County's X,Y,Z, are all sending more workers to County M, than to other counties, so these four counties should be in one region.  Ideally, there would also be a way to give greater weaight to short distances, since it's unlikely that large numbers of workers are commuting 200 miles each way each day. 

Any suggestions for how to create regions out of these counties? I don't have spatial analyst on my machine, so am directing a co-worker in another state that has the extension on his machine. 

Previous efforts, such as Hot Spot Analysis, or simple summing for each county, have highlighted counties where a lot of workers are going to, but haven't been able to illuminate the interaction between the counties.  

Thanks,
Tom

Outcomes