Data mining software is widely used in a wide array of data analysis from different perspectives like travel patterns, phone calls, online behavior and health records among others. Researchers from the University of California, Berkeley and Brown University used yearbook photos to reveal how smiling evolved in the past 100 years.
Shiry Ginosar and her colleagues developed a data mining technique using old yearbook photos to make a historical analysis of how the smile evolved through the years. Ginosar and her team downloaded 150,000 American high school yearbook photos and removed those that were not face-on portraits. The team was left with 37,921 frontal-facing photos from at least 800 yearbooks across 26 states.
The researchers grouped the portraits according to gender and the decade these were taken. The photos per decade were then merged to come up with an 'average' face and look for that period. They found that one of the most striking things that evolved was the smile.
Ginosar and her colleagues devised a simple lip-curvature metric and applied it to their dataset. According to their analysis, as time went by the smile became broader.
In the 1900s, people smiled less in photographs because they were expected to hold still, just like how they would pose for a portrait painting. It also had something to do with how society viewed beauty then.
"Etiquette and beauty standards dictated that the mouth be kept small - resulting in an instruction to 'say prunes' (rather than 'cheese') when a photograph was being taken. These days we take for granted that we should smile when our picture is being taken," Ginosar said.
As photography became more widespread in the 20th century and having portraits taken became popular, people began to smile more. In today's photos, smiles are bigger and more prominent. The study also revealed that women smiled more than men. The data mining research also highlights the evolution of hairstyle and fashion trends in the last century.
However, the researchers presented limitations of their study and acknowledged that the data set was biased. They said that the number of Americans graduating from high school increased by at least 40 percent between 1900 and 1960. Furthermore, African-American students were not represented until the middle of the 20th century.
Nevertheless, the authors attested that the findings "provide us with a unique opportunity to observe how styles and portrait-posing habits change over time in a restricted, fixed visual framework."
"We demonstrated the use of various techniques for mining visual patterns and trends in the data that significantly decrease the time and effort needed to arrive at the type of conclusions often researched in the humanities," the authors concluded in their study.