Characterizing the personality of Twitter users based on their timeline information

Abstract

Personality is a set of characteristics that differentiate a person from others. It can be identified by the words that people use in conversations or in publications that they do in social networks. Most existing work focuses on personality prediction analyzing English texts. In this study we analyzed publications of the Portuguese users of the social network Twitter. Taking into account the difficulties in sentiment classification that can be caused by the 140 character limit imposed on tweets, we decided to use different features and methods such as the quantity of followers, friends, locations, publication times, etc. to get a more precise picture of a personality. In this paper, we present methods by which the personality of a user can be predicted without any effort from the Twitter users. The personality can be accurately predicted through the publicly available information on Twitter profiles.