Proceedings of the IE 2019 International Conference www.conferenceie.ase.ro 285 THE PROFILE OF SOCIAL MEDIA USERS IN ROMANIA: INDIVIDUAL CHARACTERISTICS AND THE NUMBER OF SOCIAL CONNECTIONS Alexandru Razvan FLOREA Bucharest University of Economic Studies, Romania alflorea@ea.com Monica ROMAN Bucharest University of Economic Studies, Romania monica.roman@csie.ase.ro Abstract. The information revolution has dramatically changed the way of human living. In this study, we explain how a powerful open source tool, like R, can be leveraged to collect and examine data from social networks. Using advanced machine learning algorithms on social media data, we provided a first snapshot of the Romanian Twitter users’ demographics, and we also validated recent social theories on the dimension of social networks, such as Dunbar’s number. Among many techniques, we used API calls to connect R with Twitter, face recognition algorithms provided by Azure, and traditional statistical tests like Chi-Squared. Our results show that Romanian Twitter users are young, mostly males, and reside in the largest share in the capital of the country, Bucharest. Moreover, the research tests and confirms the Dunbar’s number, showing that physical appearance, like smiling in Twitter profile pictures, is positively associated with the number of the social connections, for both males and females. Keywords: social behavior, data collection, computer programs, social media, R, face recognition JEL classification: J11, C8, L86 DOI: 10.12948/ie2019.04.21 1. Introduction Over the past four decades, computers have reshaped the way we communicate, learn, meet each other, and many other elements of our day-to-day life. The technology redefined individuals' lifestyle so dramatically that life without computers becomes harder to imagine with every day. The growth of technology impacted all fields, including social activities. More and more people use virtual environments for social interactions such as communication, learning, meeting new people, getting the news, etc. The interest in using datasets from the online platforms to study social activities and trends increases every year. There are numerous researches done on the US social behavior, such as human mobility, happiness, cultural understanding or health status [1] [2] [3]. However, there are less studies focused on understanding the population of social media users and their characteristics in less developed countries, such as Romania. Back in the 1992, Robin Dunbar predicted that human beings cannot maintain more than 150 stable relationships [4] The theory was validated by numerous researches ref, and scientists took the theory to the next level, showing that laughter is highly correlated with the social bond of an individual.