blog

Don’t Fear the Fin – Unless You’re Taking a Selfie

07/13/2017 by Chris St. Jeor Data Communication

Shark Week. The best week of the entire year. Why do we love it so? Is it the extensive description of blue sharks’ social patterns? The scientific explanation of the effect of water salinity on the health of tiger sharks? The lengthy summary of the whale sharks’ migration? Of course not. We are drawn to Shark Week year after year for one reason only: our morbid fascination with shark attacks. 

This very fascination led my colleagues and I to do our own pre-Shark Week homework. If Shark Week has taught us anything, it is to be anxious while hanging out in the water just off the beach—we all know that surfers and swimmers are the most common victims of shark attacks. Are these nerves justified, we wondered? And should we worry when taking part in other ocean activities? In an attempt to determine whether certain ocean activities lead to deadlier shark attacks than others, we used SAS Viya to analyze data from sharkattackdata.com detailing the events that led up to more than 5,000 shark attacks worldwide.

Initial Exploration of Shark Attacks

Using SAS Viya’s Visual Analytics 8.1’s text analytics capabilities, we first identified the most frequent words and phrases describing the activity of the swimmer at the time of the attack. Viya’s default settings allow us to quickly see the most frequent activities: the larger the word, the higher the frequency of the term or phrase. As we theorized, surfing and swimming were the two most common terms. While all shark attacks should be understood and taken seriously (except that one where a person was bitten by a shark because the swimmer attempted to grab the shark for a selfie—yes, that happened) we needed to focus our work on identifying the activities which correlate most strongly with fatal attacks.

By subsetting the data to look only at fatal attacks we see a somewhat different story. While surfing was the most common activity of all shark attacks, it accounts for a smaller proportion of fatal shark attacks. We see other types of activities accounting for higher frequencies when viewing the data that represent fatal attacks.

Quantifying Dangerous Activities

To better understand the correlation between activity and severity of the shark attack, we used SAS Viya’s point-and-click interface to run a logistic regression on the main activity types. Looking at the chart below we see that if you are attacked by a shark while surfing, your chances of survival are much higher than other shark attack victims’—3.15 times higher, to be precise. And surprisingly, the people trying to touch or play with a shark (we’re talking about you, selfie guy) were the second most likely to survive a shark attack. The attacks of people trying to touch or play with a shark were 1.8 times less likely to be fatal. Inversely, attacks of people who had fallen from a vessel into the water were 4.5 times more likely to be fatal, while people attacked by a shark while stranded at sea were 8.3 time more likely to die from the encounter.

We will leave the theorizing for the reason behind these behaviors to the Shark Week experts. In the meantime, however, we hope you can use the information to breathe a little easier the next time you hang out at the beach. This quick analysis was made possible through the use of SAS Viya’s most basic functions. Whether entering the early exploratory phase with your data or working to build complicated models, SAS Viya is an easy to use and powerful software. Anyone interested in learning how Zencos can help your business with your next project can click on the “Contact Us” link below. We would be glad to hear from you. Happy Shark Week!

Image credit: Thanks Frontierofficial for the shark selfie.


We help you set the right course, develop the right solution, and transfer the right knowledge. Put our experts to work on your next Analytics project.

 …..