There could be some truth to the stereotype that those who voted for Trump last year drive pickup trucks, while Clinton voters prefer sedans. Or is this just fake news? SFGate.com has learned about a new Stanford University-developed algorithm that utilized a database of some 200 cars in US cities. The researchers behind the algorithm designed it to then pool millions of images from Google Street View, and then run it all against voting and demographic data.

The researchers, who published their findings in Proceedings of the National Academy of Sciences, say their results confirm what many already suspected."If there are more sedans, (a neighborhood) probably voted Democrat (88 percent chance), and if there are more pickup trucks, it probably voted Republican (82 percent chance)," the study found, which used voting data from the 2016 presidential election. Researchers added the margin of error of the algorithm's predictions was small. Interestingly, this algorithm can also be utilized to conduct demographic surveys or a census. For example, the research paper claims the American Community Survey, done door-to-door annually, costs taxpayers $250 million.

In other words, the algorithm can be used alongside traditional information gathering methods and be applied to find additional information on, for example, the level of physical activity of a neighborhood's occupants. The types of cars people own is just one data point that be analyzed. "While we used cars in this study, what we wanted to show was that such work is possible using publicly available images and computer vision," said the paper's lead author.