A Tufts University study analyzed about 24,000 Facebook posts on the Tufts Confessions page dating back to late 2013 and organized them into 13 topic groups. The site enables students to anonymously post messages about student life.
Soubhik Barari, author of the study and a Tufts computer science student, ran Facebook posts through a natural-language processing program he wrote in Python and separated the post contents into topical groups using Latent Dirichlet allocation. Barari also used topic modeling to identify trigrams, or keyword triplets that roughly correspond to a topic label.
The study found Tufts students felt lonely most frequently, in 22 percent of the sampled posts. Keywords associated with loneliness include "want," "sometimes," or "talk," Barari found. "It's the sum of the entire neighborhood that constructs an interpretable basis when our model converges and gives meaning to a topic," he says.
Barari believes similar studies can be expanded to other campus communities to analyze what is critically different between them.
View Full Article
Abstracts Copyright © 2015 Information Inc., Bethesda, Maryland, USA