LOGAN, Utah — While phones and social media can already pick out the faces in photos, a new study could show researchers actual places liked by users.
That information is something parks and tourism boards can use for research.
The study from Utah State's Institute of Outdoor Recreation and Tourism took 9,400 posted social photos and let Google’s cloud vision algorithm analyze them. The computers agreed with the humans 78% of the time and did well telling what was water, structures, or agriculture, but not so well on forests.
People in the study preferred similar landscapes to those who did not use social media.
At the Cache Valley Visitors Bureau, they already target their advertising
"We can really drill down details to people who are of a certain age group a certain economic background and what they like to do for fun," said Julie Hollist Terrill, director of the bureau.
And if they know just what people like, they could target even more.
"I really don't know that this information is going to make me choose between 1, 2 or 3 different specific images, but it might make me say, 'Oh, people in this group really appreciate agricultural landscape,' so for a particular ad I might choose to use something like that as opposed to a lake."
The researchers didn’t make any serious recommendations, they just told the outdor tourism industry these tools are available.