Can we predict eat-out preference of a person from tweets?

Abstract

Twitter, a microblogging site, has become a major platform of communication of users on the web. Recently, location based social networking sites such as Foursquare have become popular which enable users to publish their visited places through check-ins. In this paper, we present a study of users' eat-out preferences who share their Foursquare restaurant check-ins through Twitter. Our study reveals a strong correlation of a user's eat-out preference and the linguistic features of her tweets, i.e., word use. Hence, our proposed model enables one to predict a user's eat-out preference from her word-use in Twitter.