Every once in a while, some interesting data is put on the Internet that organization studies just eat up. Back in 2004, for example, the Enron e-mail dataset provided a unique chance to study the downfall of a company as brought on by its accounting malpractice (a.k.a. fraud).
This week marks a similarly interesting case, the leak of 76,911 secret US military reports by WikiLeaks, now known as the Kabul War Diary. The data has already been taking apart by The New York Times, The Guardian, and Der Spiegel, but scientist (e.g., Drew Conway over at Zero Intelligence Agents) are only beginning to engulf themselves in the data.
With a particular interest in the communicative constitution of organization, I dove into the data myself and dug up what I deem to be the organization of war. Take a look at the below graph.
The 200+ vertices represent military units and the 600+ edges represent communication between them. A short note, the reports feature more than 1,500 units, but I got rid of the isolates. Moreover, I assume communication to take place between units if these are at the same time in the same place. All of this information is in the reports, which “need to answer the critical questions: Who, When, Where, What, With whom, by what Means and Why” (WikiLeaks, 2010).
In addition to the graph, I pulled some quick network measures out of the magic R hat:
|A SIGACTS MANAGER||0.743119266||0.598042676||0.03356428|
|DRUID – ISAF||0.165137615||0.048233629||0.032727819|
|TF PALADIN LNO||0.128440367||0.045990081||0.032826382|
|TF MTN Warrior SIGACT Manager||0.091743119||0.039942079||0.032708177|
|TF East JOC Watch||0.073394495||0.034323879||0.032722906|
As you can see, the two central vertices are A SIGACTS MANAGER and CJTF-82, which are some managerial unit for significant activities (i.e., SIGACTS) and the Combined Joint Task Force 82. Both are important military units as their communication with others provides the proverbial glue to the organizational network.
My research is far from finished, but right now this initial shot at the data is just for play. If you’re interested in the edge list that I generated from the reports, drop me a line, I’d be happy to share.