Spying on the Spy Planes - Part Two

Some of the most memorable interactive maps of 2016 were the ones created by Buzzfeed to illustrate their Spies in the Skies investigation into the use of spy planes by the FBI and the DHS.

The FBI and the Department of Homeland Security make daily flights over American cities in spy planes equipped with high-resolution video cameras and cell phone tracking equipment. In their investigation into the use of spy planes Buzzfeed used data from Flightradar24 to identify around 200 federal aircraft. Using the Flightradar24 data Buzzfeed was then able to map out some of the flights undertaken by these federal planes as they spy on American cities.

Now Buzzfeed has also used this data to train a neural network to search for other FBI, DHS, police and military spy planes. The Buzzfeed neural network was fed data from the spy planes identified in their initial investigation and from normal commercial flights. The neural network was then tasked with identifying distinctive flight patterns found in the spy plane flight paths and not in the flight paths from the other planes. The Buzzfeed algorithm identified patterns (such as the tight turning rates of spy planes circling areas of interest) which could be used to identify other spy planes.

Once the neural network had learnt how to identify spy planes from flight path data Buzzfeed set it loose on the flight-tracking data compiled by Flightradar24.

You can learn more about the project in BuzzFeed News Trained A Computer To Search For Hidden Spy Planes. This Is What We Found. The article includes information on some of the spy planes identified by the Buzzfeed algorithm and static maps showing where these planes have been spying.