BullyBlocker: Identification of Cyberbullying in Facebook (ASUF 30005880)

Project: Research project

Project Details


BullyBlocker: Identification of Cyberbullying in Facebook (ASUF 30005880) BullyBlocker: Identification of Cyberbullying in Facebook Beyond the Bully To prevent cases of adolescents being cyberbullied on Facebook, BullyBlocker alerts parents of potential cases where their child may be a victim of this form of online aggression. Indeed, over half of adolescents and teens have been bullied online, and well over half of the victims do not tell their parents when cyberbullying occurs. Addressing this epidemic, BullyBlocker gauges atrisk adolescents who are most emotionally vulnerable by using a combination of computer science coupled with psychology research. After being granted various permissions, BullyBlocker extracts information from the childs Facebook data to be analyzed for signs of cyberbullying. This includes hurtful comments, embarrassing photos, and other methods of virtual harassment. The results are then converted into numerical factors and plugged into computer algorithms that compute the Bullying Rank, determining how well this child fits the profile of a cyberbullying victim. Moreover, one of our design goals is the inclusion of mechanisms that enable improving the accuracy of our cyberbullying identification model. Friend or Foe? Without BullyBlocker, the best way for a parent to monitor their children on Facebook is by either logging into their accounts or creating an account and friending them. Despite taking these precautions, a parent is still limited to what they can find: its easy to miss or overlook potential red flags when trying to keep up with the demanding task of watching their childrens Facebook interactions, as either a moderator or a friend. Manually monitoring their accounts may be the only obstacle preventing a parent from identifying a case of virtual harassment. BullyBlocker overcomes this limitation. In fact, BullyBlocker doesnt even require a parent to have a Facebook account in the first place. Of course, the issue of the adolescents privacy is at question. Rest assured, that although our application requires access to the childs Facebook interactions, the information and conversations we identify that include online aggression remain anonymous, thus protecting the privacy of the child. The parent is only notified of alarming aggregated statistics and a summarizing numerical risk value the Bullying Rank communicating to the parent only the potential degree at which their child may be cyberbullied. A parent will be neither a friend nor a foe to their child. Thanks to BullyBlocker a parent can remain a parent. Towards Safer Social Networking An ongoing project, the model weve designed to identify cyberbullying builds on previous research findings in the area of cyberbullying in adolescents. Consequently, weve designed our model to identify the warning signs and the states of vulnerability that can be identified from available data. Previous literature influences what characteristics are considered in the identification of cyberbullying as well as the probability of bullying within identifiable groups, where victims are typically children on the fringe of various peer groups, e.g., newcomers. In addition to our first model, our team is currently in the process of implementing an initial version of the application. After reaching this milestone, we will evaluate the accuracy of the model and explore mechanisms to improve it, integrating parental feedback and machine learning techniques.
Effective start/end date1/15/149/15/15


  • Arizona State University Foundation (ASUF): $3,471.00


Explore the research topics touched on by this project. These labels are generated based on the underlying awards/grants. Together they form a unique fingerprint.