Authors: Paul Jung*, University of North Carolina - Charlotte, Ran Tao, University of South Florida, Jean-Claude Thill, University of North Carolina at Charlotte
Topics: Spatial Analysis & Modeling
Keywords: migration, American Community Survey, regionalization problem
Session Type: Paper
Start / End Time: 9:55 AM / 11:35 AM
Room: Roosevelt 4, Marriott, Exhibition Level
Presentation File: No File Uploaded
The migration flow estimates in the American Community Survey (ACS) have been used as a primary data source in migration studies. U.S. Census now publishes annual estimates of the migration flow data in county-to-county scales, which enables deep understanding of patterns and structures of regional migration in intercensal years with abundant information disaggregated by ethnicity, gender, income and education attainment. However, individual disaggregated migration flow estimates of ACS are not sufficiently reliable to use in statistical analyses and the naïve use causes misunderstanding of the migration patterns and their structure. The risk of using high uncertain flow data like the migration estimates in ACS is still unknown. Rather than using uncertain disaggregated flow data, we develop a spatial flow aggregation method to produce more accurate migration flow data. We adopt a new geocomputation technique, a multidirectional optimum ecotope-based algorithm on the flow data (flowAMOEBA), in flow-based p-regionalization problem to reduce the uncertainty of the migration flow estimates and identify anomalous spatial interactions by the level of the uncertainty measured by CVs. The algorithm aggregates adjacent origin and destination nodes where the corresponding flows have high level of CVs, and generates the set of migration flow networks with reduced number of origin-destination blocks which the user can set the level of the uncertainty by parametrizing the cutoff of CV. Despite losing higher geographical resolution, the application of flowAMOEBA provides a more reliable migration flow map with higher accuracy in the number of migration size, directions, and locations of origins and destinations.