| dc.description.abstract | Social media influences to a great extent everyday life of millions of people, and therefore the understanding of certain patterns relating to online platforms and communities, and the ability to model and predict them, can bring substantial profits. Although many researchers already have invested a significant effort into studying this domain, many problems remain open. Most of the previous works focused on problems related to data consumption, leaving the producer’s side unchartered territory. For example, very little attention was dedicated to the understanding of the impact of motivational mechanisms on online user’s behavior. Furthermore, until recently the temporal aspect of social media was often overlooked and has not received proper attention. Finally, researchers previously put a strong emphasis on studying a set of the biggest platforms such as Facebook or Twitter, while many other important social websites remained unexplored.
This thesis addresses the above gaps in the following way. First, it presents an extensive study of the social platforms and online communities organized around food and cooking. In particular, the content production in these platforms and temporal dynamics of the communities and biases that impact user preferences are thoroughly investigated. Several of the discovered patterns were employed in practical applications. The second part of the thesis is devoted to studying incentivizing mechanisms in social media. A set of methods that are able to validate the impact of digital badges on users’ behavior are introduced, and an explanatory study of conditions influencing badges’ efficacy is presented. Finally, in the last part of the thesis, bursty patterns in social media and the problem of efficient matching of bursty series is studied. | nb_NO |
