A method for analysis of human temperament in contrast to social network data

Abstract

Currently, with the growth of the use of social networks, the possibilities of studies on social relationships and interactions have grown significantly. Understanding how users express their feelings and manifest their temperaments in social networks can be a step towards anticipating psychological disorders. Instagram has billions of users and is among the most used social networks today. However, it is still little explored as a source of study for human temperament. This work aims to analyze the relationships between users’ temperament and their data collected from the social network Instagram. For the analysis of textual data, two sentiment classification strategies are proposed. The sentiment classification results were satisfactory, with accuracy above 80% in three different databases. In order to analyze the relationship between the temperaments and social network data, statistical tests are used. Each user is represented by their positive and negative captions, the use of emojis in their posts, and the number of likes in their posts. Users of the same temperament are contrasted with users of other temperaments. The results indicate that depressed users post more captions with positive sentiment than hyperthymic, angry and worried users. Anxious users have more likes than depressed, hyperthymic, angry and worried users, and finally, anxious users use more emojis in Instagram captions than depressed and angry users.