Living in a world driven by social media, we are constantly surrounded by other people and their opinions. While the internet has connected people from all over the world and allowed them to collaborate and share their knowledge, it has also led to an uptick in crime.

Cyberbullying and online harassment are just two of the many problems that currently plague the internet.
A recent study showed that of the 6000 children between the ages of 10-18, more than 50% had experienced cyberbullying of some kind. They are also 50% more likely to commit suicide post an incident of bullying or sexual harassment. (Source: https://enough.org/stats_cyberbullying)
As young students ourselves, we have seen the kind of hatred people online are capable of firsthand. This is why we built CyberFence.

** CyberFence is designed with the intention of being a resource that can bridge the gap between law enforcement agencies and social media corporations like Twitter by using publicly available tweets to flag people who are repeat offenders. **

CyberFence allows real time social media monitoring to detect bullying and abuse. It scrapes tweets from Twitter using keywords that represent hate speech or bullying and obtains a list of users and their tweets for a 24 hour period that might be actively engaging in cyberbullying and harassment. We generated from this list of keywords from our hatespeech and cyberbullying datasets using YAKE. We then pass this data to a finetuned NLP model (BERT) that predicts whether the tweet is cyberbullying/hatespeech or not. We then process this dataset to look for any repeat offenders. If flagged we would pass this information directly to law enforcement agencies to take action.

** _Our main goal with CyberFence is to help curb cyberbullying in whatever small way we can. If we succeed in making even one individuals life better we would have succeeded in our goal. _**