Software Engineering Institute | Carnegie Mellon University
Software Engineering Institute | Carnegie Mellon University

Edge Analytics: Mining Data Streams for Intelligence

Warfighters and emergency first responders frequently operate in environments where personnel are required to quickly comprehend and react to rapidly-changing situations. The Edge Analytics system offers near real-time situational awareness (seconds to minutes) by analyzing social media and other high-velocity sensor data streams to provide actionable intelligence, trends, and summaries.

Our goal is to bring real-time analysis of data to tactical and emergency personnel by

  • applying approximation and other strategies that allow the system to meet near real-time requirements 
  • leveraging contextual clues from the local environment where they are operating 
  • providing controls (e.g., filters) that reduce resource consumption and data volumes, increase the accuracy of analysis, and tailor the output to situational needs

 The Edge Analytics System

  • performs macro trend analysis (sentiment, topic, entity and location) on data slices
  • analyzes social networks in real time to identify network structure and metrics 
  • supports interactive visualizations to allow operators to understand and digest high volumes of fast-moving data

Our research is dedicated to overcoming several challenges: large data volumes (e.g., 500 million tweets per day);time sensitivity –particularly in tactical or crisis environments;data veracity - establishing trust in data, and multi-sensor fusion –particularly open source with other forms of intelligence.