Partnering with 9-1-1 to Save Lives with Social Media

by Hfbtech

Share now:

Social media applications are becoming an increasingly important tool during emergencies. During recent emergency situations such as Hurricane Harvey and an active shooter in Maryland, social media applications have been used as a help-seeking medium when dialing 911 was not feasible or desirable. For instance, during Hurricane Harvey, 911 centers were overwhelmed with 10 times the normal call volume, and Public Safety Answering Points (PSAPs) were unable to answer all of the incoming calls. Instead, many citizens went to Twitter, Facebook, and other popular applications to post their requests for help, making it possible for first responders to act. 

During the shooting at the Capital Gazette in Annapolis, one victim quietly requested help over Twitter in the hope of receiving emergency assistance without dialing 911. 911 centers and first responders are aware of this trend, and many believe there is an opportunity to improve emergency response and save lives using these social media platforms. “We have a moral responsibility to answer the call, no matter how it is received,” stated Jim Lake, the PSAP Director in Charleston County.

Examples of emergency data shared on social media:

tweet 2

tweet 1

Understanding Social Media Use in the 9-1-1 Context

In recent months, RapidSOS has partnered with the Pennsylvania State University (PSU) Department of Information Sciences and Technology, Mission Critical Partners (MCP), and Charleston County 911 to conduct research on how social media applications can be used in an emergency situation and to build a proof-of-concept tool for surfacing requests for help on social media. The team began its research work in the spring and early summer by conducting observational research and interviews with 911 call-takers and dispatchers as well as leaders and supervisors within the Charleston County PSAP.

The collective goal of the research was to understand the situations in which social media might be used and the type of information that is needed to safely and effectively dispatch first responders. Most importantly, the team wanted to understand the context in which call-takers and dispatchers operate to understand how information flows into and through the PSAP.

Social media pilot screenshot

Following the visit, RapidSOS worked with PSU and MCP to incorporate the research into initial use cases and requirements for a proof-of-concept to quickly be built and tested in simulations with the call-takers and dispatchers in Charleston County. Using publicly available social media data, the team was able to quickly build a tool that could ingest unstructured social media data, filter it using various open-source machine learning and analytics techniques, structure it, and surface content that might be related to ongoing emergencies. Researchers were also able to demonstrate how emergency information on social media could be integrated directly into Computer-Aided Dispatch (CAD) systems to generate an informed emergency response.

The team returned to Charleston in early August to conduct a simulation using various emergency scenarios. Call-takers and dispatchers were assisted by a designated communications specialist who was able to share information about emergencies implementing social media data. The response was overwhelmingly positive, and we were able to gather user feedback on how social media data can benefit telecommunicators inside the PSAP.

Given the positive response, the team remains focused on developing technology to integrate social media data into PSAP workflows. Working with Charleston County offered the opportunity to implement user feedback from the simulation to apply towards a tool to use in emergencies. Additionally, we are paying close attention and strictly adhering to data privacy rules and regulations.

We want to ensure that we build a tool that minimizes data collection and protects individual privacy even when assistance is requested in an emergency. We will accomplish this by fine-tuning the machine learning algorithms to reduce noise in the data and never capturing excess data. We’re excited to continue working with the 9-1-1 community on developing a solution to provide telecommunicators with additional life-saving data.