The other day, while casually browsing for interesting geospatial datasets (as you do), I came across an enticing trove of map-worthy material: the U.S. Department of Defense’s recently released Theater History of Operations data, a comprehensive database of historical U.S. bombing missions from World War I through the Vietnam War. As a kid, I was fascinated with military aviation—I read the “Encyclopedia of 20th Century Air Warfare” cover to cover nine times, and built two dozen scale airplane models, hanging them from my ceiling with fishing line—and so I was naturally eager to take a crack at the data.
I decided to begin by mapping the Vietnam War dataset—the largest of the bunch, by some distance. After downloading the Vietnam War dataset as a CSV (which tipped the scales at a hefty 1.6GB), I brought it into R for some basic cleanup and analyses. Specifically, I standardized the dates, some of which had been entered improperly; I removed unwanted records, including non-combat missions and missions with invalid target coordinates; and I aggregated the number of bombing and ground assault missions per month, to visualize in a column chart. The original dataset contained about 4.6 million unique missions; I managed to whittle this down to about 3.1 million missions. Much better!
Next, I brought my refined dataset into QGIS (avert thine eyes, my Esri colleagues! I have to run ArcGIS Pro in a VM on my MacBook, and it kept choking on the layer) along with some reference layers from Natural Earth, and applied some simple styling. Each mission is represented by a single, nearly transparent point, so that isolated missions are hardly visible, while areas of intensive saturation bombing are nearly opaque. Finally, I brought the map into Illustrator, where I added the chart of monthly missions and the annotations.
This map barely scratches the surface of the Vietnam War dataset, and I intend to explore it further—along with the other THOR datasets—as and when time allows.