As the world focuses on today’s U.S. presidential election, with intense competition between Republican Donald Trump and Democrat Kamala Harris, a team of researchers from New York University, led by Moroccan computer science professor Anasse Bari, has developed a new method to analyze voting intentions and predict election outcomes.
The method uses artificial intelligence and data aggregation to track voters’ online search behavior, specifically searches for candidate-related items (such as hats, shirts, or banners). The research team believes this approach can complement traditional polling methods.
In an interview with CNN, Barry explained, “If someone searches for a banner, hat, shirt, or flag related to a candidate, it indicates a strong likelihood of supporting that candidate. Similarly, searching for terms from a presidential debate suggests interest in the debate itself.”
Barry highlighted the growing importance of social media and blogs in expressing voter intentions, stressing the need for innovative methods to measure public sentiment beyond traditional polls.
The team’s research identified key phrases people use when searching for candidate-related products and developed indicators to track these behaviors. They also employed sentiment analysis using AI algorithms to assess voter preferences.
Barry noted that the method can complement traditional polling by offering a deeper understanding of voter sentiment and political leanings. The study found, for example, that Trump’s hats are particularly popular in several states, while Harris performs better with her campaign symbols.
This post was originally published on here
