The Industry Space

This is an R Markdown document intended to show the latest network visualizations produced in the Workforce of the Future Project.

LetÂ´s beging with the industry space. This visualization shows NAICS 4 digits industries (version 2003) joined to eachother according to their probability of co-location across Metropolitan Statistical Areas. The location of each industry is based in the patterns of location quotioent of the employment level of each industry in each city recorded in the CBP.

The industry space II

Alternatively, the industry space can be based in a proximity matrix that shows how related are two given industries according to the intenisy in which they use workforce inputs.

The Skills Space

The logic under which the Industry/product space is drawn- as well as the network anaylis methods that are pontentially appliable to it- can bring ligth into other topics, such as the skill requirements of different jobs.

Following Alabdulkareem et al (2018), we present a set of 161 skills from the ONET dataset (They are a mix of Skills, work activities, abilities and knoledge) connected to eachother according the minimun conditional probability of skill X being used by occupation O given that skill Y is also used by O.

After aplying the Louvain algorithm for weighted community detection over a network made of 161 nodes (skills) and 161x161 edges (porximities), two clusters emerge: One contains cognitive skills (blue) and the other physicall skills.

Next, we keep the higher 20% of proximities along the higher proximity of each node, so we can visualize the network easily.

The Occupations Space

Now we show occupations (SOC10 clasification) interconnected according to the similarities of the intensity whit which they use each skill registered in the ONET 23.0 data set. (WORK IN PROGRESS: the occuspace has 685 and 46855 edges, which results in a network three times bigger than the industry space).