Lying between the Italian peninsula and the northwestern Balkan nations, and separated from the Mediterranean Sea by the Strait of Otranto, the Adriatic Sea is largely cut off from global-scale ocean circulation patterns. As a whole, the Adriatic has a permanent counterclockwise circulation, but in the shallow northern reaches, surface currents vary dramatically over short spans of time, with potentially dangerous consequences for maritime activity. To understand the primary drivers of the fluctuating surface currents, Mihanović et al. perform self-organizing map (SOM) analysis—an emergent computational technique in oceanographic research—on data provided by three high-frequency radar stations operating in the region. As a neural network technique, SOM analysis uses complex mathematical algorithms to train computers to pull patterns from jumbles of data, reducing complex multidimensional observations into simple visual maps. The approach is meant to emulate the learning abilities of biological brains.