NYC First Responders Join Forces with Columbia Faculty

Nov 05 2018 | By Holly Evarts | Image Courtesy of Columbia Research

Making sense out of the chaos of a disaster zone is key to preventing casualties and jumpstarting recovery. As sensing technology evolves, data science holds the potential to generate life-saving insights, but how to best connect the dots between data in the field and the first responders who must quickly interpret the flow of information has proven a consistent challenge.

In late September, Computer Science Professor Dan Rubenstein convened a workshop to address just these issues. Drawing together more than 35 attendees from the New York City Fire Department, Police Department, and Emergency Management, along with professors from Columbia University and John Jay College of Criminal Justice (JJC), the event served as an introductory bridge between first responders and the academic community. The workshop on Human-Centric Automation of Information, Coordination and Analytics (ERICA) focused on exploring how massive, evolving sets of data can be transformed into useful information to support real-time decision making and response coordination and to show how data analytics can be leveraged for improved preparedness and response.

“The workshop allowed NYC first responders to see how research at Columbia and John Jay might be useful to the first responder community, and allowed faculty to better understand the concerns of first responders to guide future directions,”  said Rubenstein, who also co-chairs Sense, Collect, and Move Data, a center at Columbia’s Data Science Institute that brings together researchers focused on the physical aspects of generating, collecting, storing, and transporting large data sets. “We are all very excited about the many collaboration opportunities ahead to make a real difference to the safely and security of New Yorkers and beyond.”

Hosted by the University and JJC on September 21, the workshop featured research talks by Computer Science Professors Lydia Chilton, Henning Schulzrinne, and Shih-Fu Chang; David Yao, professor of industrial engineering and operations research; Norman Groner, an assistant professor at JJC, and NYFD Deputy Chief Michael Meyers that led to lively discussions. The participants explored how to better understand the interactions among computing, data analysis, resource allocation, and real-time responses. The discussions offered the researchers a closer look at real-world constraints and spurred conversations about how to design computing systems to aid first responders, and how to address the gaps in converting data into useful analytics that can contribute to effective real-world coordination.

ERICA is the first workshop Columbia has hosted to generate collaborative discussion among academic scientists and engineers and emergency first responder community members. Faculty focused on helping to develop new first responder technology, such as integrating information flows, improving resource allocation, artificial and virtual reality applications, machine learning, identifying role responsibilities, and developing communication infrastructures. Discussions centered around identifying operational challenges and defining shared vocabulary, assumptions, and priorities.

All agreed that the goal is to turn information into intelligence that can be readily used. But many obstacles currently stand in the way of seamless coordination of first responders in the field, due to disconnects between data analytics, data transmission, and coordination that continue to slow response time and increase risk. First responders were concerned with a number of issues:

  • Improving emergency responder systems, helping to connect networks and to relay information efficiently and effectively among and across organizational structures.
  • Developing a common operating picture of real-time data in an incident to be sent out to all agencies and organizations involved.
  • Synthesizing and prioritizing massive amounts of data.
  • Minimizing the amount of incomplete data/gathering data from community/rumor control.
  • Developing a way to combine geospatial data and social media mining to create data streams to track and manage incidents.
  • Improving and managing technological advances in research systems’ efficient real-time facial, action, object, and group recognition technology and traffic detection technology across surveillance platforms.
  • Effectively incorporating human limitations in interacting with new technology to prevent technology from being rejected for more traditional and inefficient methods.
  • Using complex, agent-based simulation models to improve responder coordination, training, and preparation; managing the loss of institutional knowledge due to employee turnover; anticipating traffic issues and patterns.

Carla Murphy, NYFD manager of EMSCAD (Emergency Medical Services Computer Aided Dispatch Programming) noted that the goal is how to get the information to the field more efficiently to help agencies determine if additional help is needed.

The collaboration offered an opportunity to discuss issues and brainstorm about solutions, and start a dialogue around the various issues that will benefit from an interdisciplinary approach.