For my first data visualization, I wanted to look at the career that Navy veterans pursue post separation. My original intention was to show this for all Navy personnel. However, using LinkedIn's "Advanced Search" capabilities, this proved to be too cumbersome to collect this data set. Additionally, refining the search to all Naval Officers proved too onerous. As a result, I used US Naval Academy graduates as a proxy for all US Naval Officers. The final result of this initial data can be viewed here.
This resulted in data sets for Submarine Officers, Surface Warfare Officers, Naval Aviation Officers, and US Marine Corps Officers. Unfortunately, there was insufficient data to do a similar set of research around SEALs, Supply Corps Officers, Seabees, etc.
I grouped the data around several key areas related to a post-military career.
- Industry: I looked at all 147 LinkedIn classified Industries, except for "Military." This was the easiest way to look at Naval Officers on LinkedIn who are no longer on Active Duty. The 147 categories did not provide as much insight, so I created my own subset of categories in order to extract higher-level takeaways. I was unable to find any official guidance on the best way to create these subgroups, but provide an overview of my grouping here.
- Function: Fortunately, LinkedIn only provides 12 different categories for Functional Roles. However, LinkedIn only shows the top 10 industries for any designated search. This means that for each service group, there were two industries not shown. For Submarine Officers, this was Consulting and Education. For Surface Warfare Officers, this was Research and Education. For the US Marine Corps, this was Business Development and Research. And for Naval Aviation, this was Finance and Research. In this case, I inserted the data as zero, knowing that the actual percentage is higher than that, but also not in the top 10 Functions for any service group.
- Size of Company: This was the simplest to obtain, as LinkedIn only provides 9 categories for Company Size, and provides this data consistently for all service branches. As a result, the data here will be the most accurate to actual LinkedIn data. For my article on LinkedIn, I divided these into just 5 categories to simplify the high-level takeaways. However, I've provided the original data within the chart.
- Length of Service: This was the most complicated to extract. In order to estimate the Time in Service for the nearly 5,000 LinkedIn profiles analyzed, I needed to use the third-party service, Amazon's Mechanical Turk (mTurk). I provided spot checks of the data results, but due to limitations in both time and budget, was not able to verify all of their work.
In order to display the data, I used the New York Time's D3 model. Special thanks to Nemil Dalal, who put together the majority of this data, and helped me as I put together the small remainder he did not complete. I also used Upwork in a few locations to help me edit this models and add them to my Wordpress website template.
The next set of data I would like to look at is how Time in Service affects Industry, Functional Role, and Size of Company of veterans. However, I wanted to provide the initial data first in order to see what additional information would be most helpful to active duty military personnel. All feedback and suggestions are welcome here.