A look at Puri's ever-changing map, as of this morning. The hotspots indicate mentions of bullying on social media. (Photo: Vertabox)
A 13-year-old from Virginia is trying to use technology and big data to fight bullying. Viraj Puri runs the website Bullyvention, an anti-bullying site that works with the U.S. House of Representatives’ Congressional Anti-Bullying Caucus. He's now building a heatmap that would track instances of bullying — in real time.
Puri uses an algorithm to search for keywords — like "bullying" and its variations — on Twitter, Facebook and other social networks. (He's hoping to partner with those sites, but no formal arrangements have yet been announced.) The data is then analyzed to determine whether somone is bullying or being bullied, and that information is plotted on a map.
Puri was inspired to start his anti-bullying campaign after watching his older brother get harrassed after school. His goal is to identify areas where bullying is most prevalent. Puri explains his mission in a release about the project:
Our goal is not only to “predict” the likelihood of the next bullying occurrence but see where it’s happening geographically, which begs the question ‘Why it is more of an epidemic in some areas of the country over others?’ New homeowners look into the quality of schools, crime, jobs etc. when they are looking to move into a new area, parents should have an indicator as to how bullying is in each county, city and school district across the United States before they move into a new area. This will put more focus on quality of life and get lawmakers to address this very important issue.
Puri is working with Dr. Xiaojin Zhu, a data expert at the University of Wisconsin, to help sift through the data. Zhu's research shows that only about four per cent of all uses of the word "bullying" online are actual attacks. Other instances include responses, anti-bullying messages or news articles. The challenge is to figure out how to filter these out and only show real bullying attacks on the map.
Puri's project is currently using a sample size of between 500 and 7,000 tweets per hour. As of now, the map is still in beta and the data hasn't yielded any conclusions.
He's hoping that as it grows he'll be able to create a more reliable index — and that the index will eventually be useful to people who get bullied.
"Even if I save just one person from committing suicide," Puri tells the BBC, "that's a big difference right there."