Deploying Social Media Data to Inform Gun Policy

Overview

This study will investigate the use of data from social media posts to measure gun-related outcomes in cities or states, and improve the data available to support rigorous gun policy research.

Status

Complete

Purpose

This project seeks to improve the data available to understand gun-related outcomes and support rigorous gun policy research through the novel use of social media data.

Approach

• Specific Aims:
  • Develop estimates of gun-related mortality and benchmark against estimates based on data from the National Vital Statistics Systems and the Centers for Disease Control.
  • Develop algorithms to predict gun ownership status.
• Use the large repository of Twitter data housed at Georgetown’s Massive Data Institute (MDI) as primary data source.
• Collect additional social media data using keywords and hashtags related to gun violence.
• Conduct a manual review of gun violence/mortality-related tweets and code each tweet as identifying a gun-related homicide, gun-related suicide, or any gun-related death.
• Using manually labelled data as training data, develop a classifier to identify tweets discussing gun-related deaths.
• Develop a second classifier that determines type of gun death.
• Develop predictive algorithms for gun ownership status from Twitter posts.
• Conduct validation testing throughout the study.

Significance

Build a foundation for understanding for how social media data may be used—alone or with other data resources—to improve gun policy research.

Published Research

• A Comparative Analysis of Classic and Deep Learning Models for Inferring Gender and Age of Twitter Users

  View research

• Analyzing the Impact of Missing Values and Selection Bias on Fairness

  View research

• Assessing Social Media Data as a Resource for Firearm Research: Analysis of Tweets Pertaining to Firearm Deaths

  View research

• Development and Assessment of a Social Media–Based Construct of Firearm Ownership: Computational Derivation and Benchmark Comparison

  View research

• Social Media Data for Firearms Research: Promise and Perils

  View research

• Social Media Data—Our Ethical Conundrum

  View research [PDF]

• Students or Mechanical Turk: Who Are the More Reliable Social Media Data Labelers?

  View research on OSF

• Text Analytic Research Portals: Supporting Large-Scale Social Science Research

  View research (subscription required)

• Traditional and Context-Specific Spam Detection in Low Resource Settings

  View research

Investigator Bio

Lisa Singh is a professor in the Department of Computer Science, and research professor at the Massive Data Institute, of Georgetown University. Dr. Singh has authored or co-authored over 70 peer-reviewed publications and book chapters related to data-centric computing. Current projects include learning from public, open-source big data to advance social science research on human behavior and opinion; identifying and quantifying noise and other forms of poor-quality information (including misinformation) on social media; developing methods and tools to better understand forced movement due to conflict; and studying privacy on the web.