Yahoo Researchers Developing Mapping Algorithm That Finds Most 'Beautiful' Routes

Taking the scenic route may be a GPS option in the future, thanks to Yahoo! Inc. (NASDAQ:YHOO) researchers.

A study published by Cornell University’s arXiv, Yahoo Labs’ Daniele Quercia and Luca Maria Aiello, along with Rossano Schifanella of the University of Torino, Italy, demonstrates how mapping algorithms could be tuned to provide directions that balance brevity with beauty.

How was this achieved?

The research team gathered images of streets in London using Google Street View and Geograph, which were then uploaded to website UrbanGems. With the website established, the team crowdsourced data from users who voted on which images looked more beautiful, quiet or happy.

That data was then used to develop algorithms which provided aesthetically pleasing walking directions.

Surprisingly the resulting routes were only 12 percent longer, according to the study. In comparison, itineraries that focused on tourism and entertainment directions were 60 percent longer than the shortest routes.

Furthering the study, the team also attempted to automate their results in Boston. Armed with 1.3 million Flickr pictures from the city, the team used the metadata available from the images to “predict beauty scores,” which were then used to build scenic routes. Fifty-four participants were then recruited to compare directions based on the paths generated by the Flickr metadata and the shortest directions. The participants rated the found the paths generated using Flickr metadata as 35 percent “more beautiful than the shortest paths.”

There are certain flaws to the study. For example, it doesn’t take into account how time of day or rush hour traffic affects a trip.

But it is an interesting and different approach to changing the tourist experience in the city.

According to MIT Technology Review, Quercia and Yahoo Labs plan to develop a mobile app based on their study in the near future.