PA Bill Number: HB2235
Title: Providing for regulation of the meat packing and food processing industry by creating facility health and safety committees in the workplace; ...
Description: Providing for regulation of the meat packing and food processing industry by creating facility health and safety committees in the workplace; ... ...
Last Action: Referred to LABOR AND INDUSTRY
Last Action Date: Apr 25, 2024
How the 'digital exhaust' of social media data can predict gun violence :: 03/15/2016
When it comes to data on shopping habits or driving, it’s all available and being constantly mined by advertisers and government offices who sell and legislate around it appropriately.
But data around gun violence? There’s almost nothing. And that had been by design – until now.
Unveiled at the SXSW technology festival, a group of data scientists and activists have demonstrated for the first time a new way to predict and study gun violence using social media. Scraping Tweets, Google searches, obituaries and local news, they’ve created a livesteam of gun-related discussion and a map of violence and geo-tagged posts. By reading this “digital exhaust” of data they hope to create “digital phenotypes” and understand how people are behaving around guns, when and why. The Gun Violence Surveillance website launches in the coming months, and those wanting to get an alert when it’s live can sign up online.
“The data’s there, or it can be there easily. The FBI has data. The police have data. But we’re forced to do this because no one is allowed to share it,” said Jessyca Dudley, program officer of the Joyce Foundation’s gun violence prevention program. “So we started looking at unconventional data sources.”
But they’re nervous that the police or FBI might use this strategy to predict individual people’s likelihood of committing violence: “Law enforcement may want to do that,” said Dudley. “We don’t.”
Since 1996, the Center for Disease Control has been almost entirely banned from studying gun deaths thanks to the Dickey Amendment restricting federal funding for gun violence research. And so Dudley and her team – Ben Reis, director of predictive medicine at Harvard Medical School; Evan DeFilippis, founder of Armed with Reason; and John Brownstein, chief innovation officer at Boston Children’s Hospital – got together to build a tool of their own. “We’re looking to expand the digital phenotype,” Brownstein said. “The digital exhaust can provide insights.”
Honing in on Austin, using the new site can zoom through different different neighborhoods reading geo-tagged tweets talking about people hearing gun shots or debating policy. It’s timely here in Austin where a gun rights activist was recently shot by her toddler.
Data is key for making legislation, they argued. When a car kills a person, the government records weather conditions, seatbelt use, drug use, and someone can search “what is the risk of an auto fatality for a woman at a stoplight in the fog?”
“It directly informs safety standards. Thanks to this data we have drunk driving laws, minimum age drinking laws,” said DeFilippis, founder of Armed With Reason, a site dedicated to providing a scientific defense of gun violence prevention.
But are people really tweeting this much about guns? According to Brownstein, more than 20% of all tweets from Idaho are related to guns. Even looking at what people search can provide insights into how many people in specific areas are buying guns or dying of gun violence.
Measuring Google searches for the word “pistol” in Idaho, for example, correlated with government data on how many people were getting background checks to buy guns.
Gun deaths in regions correlated to the Google searches for “ammo”.
