Skip to main content Link Menu Expand (external link) Document Search Copy Copied

Goal

The goal of this data science effort is to mitigate cyberbullying for children on social media by identifying vulnerable young people. Particularly, this effort will focus on U.S. teens on Instagram.

Problem Overview

Cyberbullying is prevalent among U.S. teens. The Pew Research Center reported that 46% report experiencing cyberbullying behavior.

76% of teens believe that social media companies looking for and deleting posts they think are bullying or harassing would help reduce harassment on social media platforms (Pew Research Center). Given only 2 options, 74% of teens say social media sites are doing only a fair/poor job addressing online harassment compared to the 25% that say the sites are doing an excellent/good job (Pew Research Center).

Given these statistics, it is paramount that social media sites take action. Therefore, this data science effort will be created with the assumption that it will be carried out internally by Instagram.

What Exactly Does Cyberbullying Look Like?

The most common type of cyberbullying are comments, rumors, and sexual remarks. 22.5% of students reported being the victims of mean comments online, 20.1% were the victims of false rumors, and 12.1% were the victims of sexual remarks (Cyberbullying Research Center).

The comments and direct messages perpetuating these types of cyberbullying are primarily text and can be identified through natural language processing methods.

Who Is Being Targetted And Where?

Age, gender, race, income, physical appearance, sexuality, and political views affect teens’ cyberbullying experiences and the degree to which they view cyberbullying as a significant issue. Generally, older girls and white teens report false rumors spreading, and teens from low-income backgrounds are twice as likely to have been physically threatened online. Black or Hispanic teens are more likely to say that cyberbullying is a significant problem.

Although our data science efforts can help us identify vulnerable individuals, understanding and acknowledging the demographic factors of victims can make our efforts more fruitful and the resulting actions more targeted.

More young people experienced cyberbullying on Instagram (42%) than any other social media platform (Enough.org). Therefore, the focus of this project will be to mitigate cyberbullying on specifically Instagram.

Sources

  • https://www.pewresearch.org/internet/2022/12/15/teens-and-cyberbullying-2022/
  • https://cyberbullying.org/2019-cyberbullying-data
  • https://enough.org/stats_cyberbullying