Most shark attacks occur near the beach…
Humans have a tendency to get focused on the minutiae of what they are doing and forget the end goal. We “miss the forest for the trees” or, as my drill sergeant used to succinctly say, “Don’t miss the war while staring at your feet.” We need to remember in this era of “big data” and “real time analytics” that just because you can describe a phenomenon that doesn’t mean your description is useful.
As a former intelligence analyst I am used to working with data to try to provide insight for commanders. An Intelligence analyst’s job is to utilize vast amounts of data and to provide answers and insight for my commander and for our maneuver units. You are trying to give the guys on the ground a tactical edge over their enemies. Battle is a zero sum game so there is no margin of error. That being said, intelligence is the art of illuminating the darkness but sometimes you end up just staring at your feet. This is my story of how I began to understand that there is a real difference between describing your environment with maps and charts and statistics and understanding your environment.
An ambush, simply, is something to be avoided, and my commander wanted our guys to be in less of them. My intelligence cell started mapping where our maneuver units had been in ambushes, what type of weapons were used, the intensity and all manner of different variables related to these events. Soon enough, a clear pattern emerged. There were indeed certain streets and certain areas where ambushes were more frequent; we referred to these areas as “hotspots.” This was presented to the commander and the maneuver units and they either changed their routes or approached the hotspot areas with greater guile or planning. This was the edge we were looking for. Or so we thought…
The ambush numbers did not decrease. Our guys were getting into just as many ambushes as before, but now the ambushes were in different areas, away from the hotspots. Once we changed our pattern the enemy changed theirs. What prolific adversary! We had used advanced computer mapping and skilled analysis to adjust our behavior and gain the tactical edge, but our enemy, despite a lack of resources, had followed suit. What was going on here? We began to look at the demographics of the areas, where the people worked, who was dominate politically etc. etc. etc. But we could find nothing to explain.
While trying to get to the bottom of this “wicked” counterinsurgency problem of why do we fight in certain areas, I started working with some of the guys in the operations cell. The “Ops Cell” tracked the movement of all the maneuver units and managed the day to day operations. From them I grabbed our units route information, the GPS plots of where our units had been in the past 30 days, and I potted them to a map. To my surprise the pattern of our movement looked exactly the same as the pattern of “hotspots.” I then plotted our unit’s movement after the hotspots were “discovered” and the new “emerging hotspot” pattern was evident.
The “hotspots” of the original analysis were simply the areas our units spent the most time. The enemy was attacking us there because that’s where we were. When our units started avoiding those areas the enemy simply attacked them in the new area. We had only “proved” that there was, indeed, a war going on and that we were involved in that war; but now we had the data and maps to prove it. We were staring at our feet.
“Because thats where the money is.” -Willie Sutton, bank robber.
Intelligence Analysis is using data to help understand your enemy. These same sorts of skills and solutions are migrating into other fields, often under the moniker of Data Science or Informatics. I am always on the lookout for “insights” in disguise; the deep staring at your feet. Often you will see researchers or media serve up an interesting study or explanation of natural phenomenon, the classic being “Most shark attacks occur near the beach…” Now I am sure they have the data to back it up but someone needs to only ask the question, “…well, where are most people in the water.” The answer, the people are near the beach. Most shark attacks on people happen where people come into contact with sharks, near the beach. It can’t be any other way, there is no alternative.
Some more of my of favorites are the perennial map of “Cities Most in Danger from Sea Level Rise.” Spoiler alert, it’s the cities in poor countries located on lowland near the sea. Insightful. As a resident of Florida my insurance company sent me a scholarly article which declared Florida as the “number one state for boat insurance claims.” The article failed to mention that Florida has the most boaters, and the most boats. People with boats make boat insurance claims. Poignant. Guess where the majority of people are located who use Twitter? If you guessed “areas of high population density,” you just might be a data scientist. Not a lot of tweets from the South Pole, or the South Pacific, or areas with no electricity. For this reason we need to give sharks Twitter accounts so that we can use analytics to see when they are near the beach.
Of course I jest. I am being obtuse. I am sure there are vagaries about when sharks are near the beach and why so many boat insurance claims happen in FL. And there were subtle differences as to why we were ambushed more in certain locations, but I believe my larger point is clear.
The initial impulse to just catalog, quantify and map is understandable. It can answer the essential questions of what, when, where and how much which is necessary. We need to keep in mind the why; not only why is a phenomenon occurring, but why do we care. For data science to be meaningful it needs to illuminate the darkness. We must make sure we are not just staring at our feet.
Sources of pictures:
SIThereAndThere. “Wthis Is Most Accurate Pic I’ve Found. • /r/pics.” Reddit. N.p., 21 Mar. 2014. Web. 24 Oct. 2016.
Nowakowski, Kelsey. “See Where Most Shark Attacks Happen in the United States.” National Geographic. National Geographic Society, 11 July 2015. Web. 24 Oct. 2016.
Star, Tom Beal Arizona Daily. “UA Climate Research: Big Stretch of US Coast at Risk of Rising Seas.” Arizona Daily Star. N.p., 24 July 2014. Web. 24 Oct. 2016.
