Labor Market Areas (LMAs) are distinctive communities of counties within a nation’s commuting network that are closely connected with one another than with other counties. The overlapping of communities within a hierarchical structure is a crucial feature of real-world networks, including commuting, yet existing methods are inadequate for modeling such overlapping because the concept is inconsistent with hierarchical ordering. Ahn et al. (2010) introduced the link community method to address this problem but did not consider weighted and directed links such as commuting flows. In this paper, we extend the link community method to accommodate directed and weighted networks using the idea that edges can be presented as vectors. We then apply our proposed method to U.S. commuting data. Results suggest that our new method reliably identifies the LMAs that we would expect to find. In our case, however, these LMAs can also overlap one another.