Alexander Fishkov, Ph.D. student Computer Science
In this post continue analyzing data on employment and migration. We are interesting in workers who reside in one county and go for work in a different one. This time we got the data from the Metropolitan and Micropolitan section of the Census.
An interesting thing to look at is the structure of worker migration by identifying geographically compact clusters of counties based on how many workers travel within them as opposed to between them. If we view this as a graph, counties will be the vertices and worker migration flows will become the edges. We set the edge weights to be the estimates of number of workers traveling in this direction. In the field of Network Science this task is called community detection, and there are existing algorithms to do this. We have chosen the algorithm by Blondel et al from their paper Fast unfolding of communities in large networks. It is implemented in Gephi – a free graph visualization and exploration platform. Below you can see a map of highly connected counties, grouped together by color.
The algorithm identified 38 larger regions or “clusters” of counties, which is less than the usual partitioning into 51 states. The first thing to mention is that all regions are geographically consistent, meaning that counties of the same group form a bounded region on the map with no members of other groups within that boundary. It is also noticeable that there are very few “enclaves” - regions that entirely lie within a single state. Such regions are only found in Alaska and South Dakota. The discovered regions mostly span over two or more states. The largest number of states spanned by a region is five, and is situated at the south coast: Louisiana, Mississippi, Alabama, Florida and Georgia.
As we have seen, this employment statistics approach suggests a division of the country that is quite different from the one we have now. Maybe it is more reasonable than the political division into states? In fact, we have not gathered enough information to state that. There are other important economical as well as geographical features to take into account.
About Alexander Fishkov
Alexander Fishkov, Ph.D. student Computer Science
Alexander is a Ph.D. student in Computer Science. He currently holds B.S. and M.S. degrees in Applied Math. He has experience working for industry major companies performing research in the fields of machine learning, data mining and natural language processing. In his free time, Alexander enjoys hiking, Nordic skiing and traveling.
Other posts by Alexander Fishkov: