Abstract:
|
Most analyses performed are concerned with obvious relationships: exactly matching Alien Numbers in different datasets, or – similarly – matching addresses, receipt numbers, or user identification numbers. However, patterns of behavior or identification of risk are generally not so obvious nor easily discovered among exact matching data. This panel discussion explores the nature of non-obvious data relationships such as temporal and spatial combinations, but more importantly extensions into how patterns can also be identified among metadata such as with concepts like centrality (i.e. eigenvectors). While discussions of non-obvious relationships are frequent in academic and industry literature, less is known about applications and complexity within uniquely immigration data; for instance, data such as class of admission, port of entry, or free-text such as long-form, narrative statements are all objects that can be transformed and/or grouped to expose how particular actors navigate the U.S. Immigration system.
|