MCP In The News: Why Leveraging Social Media Data Is Important

This artcle, authored by Dave Sehnert from Mission Critical Partners, originally appeared in APCO's Public Safety Communications May/June 2020 issue. The online version can be viewed here.

One of the first instances of social media being leveraged on a large scale to enhance public safety and emergency response occurred in 2012 in conjunction with the Super Bowl played at Lucas Oil Stadium in Indianapolis. A 2,800-foot social-media center was established near the stadium, where 50 volunteers monitored the chatter. Platforms monitored included Facebook, Twitter, MySpace and YouTube.

Off-the-shelf software monitored more than 65 million postings, tweets and blogs over a six-week period, using 500 keywords that had been developed for the event. To make sense of it all, students at Ball State University in nearby Muncie, Indiana, developed algorithms that divided the chatter into positive and negative categories; a color-coding scheme was developed to enable volunteers to discern between them quickly and easily.

While the effort uncovered no events that required a significant emergency response, it gave public safety officials a greater comfort level given that about 1 million people jammed the downtown area where the stadium is located for a variety of fan-related events in the days leading up to the big game.

Since then, the use of social media by law-enforcement agencies has proliferated. A survey conducted by the International Association of Chiefs of Police in 2016 reportedly indicated that about three-quarters of police officers use social media to investigate crimes that have been committed or to gather intelligence in the hope of preventing them. Social media also is being used to communicate more effectively with citizens, and to gain insights into their thinking regarding emergency response services and those who provide them. Another study conducted by the Brennan Center for Justice at New York University School of Law indicated that 158 jurisdictions across the country spent $5.16 million on social-media-monitoring software from 2013-2016.

Law enforcement monitoring of social media also has become more sophisticated. For example, some agencies use Twitter Maps to enhance situational awareness. The platform enables public safety personnel to see all tweets related to a specific event over the previous six hours, simply by typing in keywords, such as “tornado” or “wildfire.” All corresponding tweets within the affected area populate on a map display. Often images are attached to the tweets, which helps officials discern the extent of the damage.

Twitscoop is another tool that has been leveraged successfully by some agencies. It is a web-based client that leverages Twitter’s application programming interface to assess what people are talking about at any given moment. The more that people are chatting about a certain topic, the larger the keyword(s) associated with that topic will appear on the screen. By typing a keyword into Twitter Map, officials can see what people are saying about the topic in their jurisdiction.

Reducing noise

Although social media represents an effective tool that enables public safety agencies to increase situational awareness—in turn enhancing emergency response and even keeping emergency responders safer—the data generated needs refinement to be actionable. The algorithms referenced earlier in this article represent the alpha, rather than the omega, of the effort. Even when using the best algorithms, there’s still a lot of noise that needs to be filtered out.

In September 2018, a pilot project that tested the use of social media data in emergency response was completed. The project’s goal was to develop a software tool that would gather and filter social media data to extract information that emergency responders would find actionable. The idea was to figure out exactly what is valuable in that data, when it is valuable, what form it needs to take to be valuable and—most importantly—how it fits into the workflow of existing emergency communications centers (ECCs), to provide responders with better situational awareness while en route to an emergency. The tool searches, filters and maps relevant information from a large dataset of real-world social media posts captured during emergencies so that telecommunicators can focus on the emergency at hand without having to dig through irrelevant information.

The following organizations participated in the project:

• Charleston County (S.C.) Consolidated 911 Center (which hosted the project)
• PSU’s College of Information Sciences’ 3C Informatics: Crisis, Community and Civic Informatics (which is led by Dr. Andrea Tapia)
• Mission Critical Partners (MCP)
• RapidSOS
• RapidDeploy

Dr. Jess Kropczynski, associate professor of information technology at the University of Cincinnati, also was a leading collaborator. Over the course of the project, social media data was filtered using a set of questions known as the “6 Ws”—who, what, when, where, why and weapon. In case the reader is wondering, “weapon” is included because 911 call-takers typically are trained to ask the caller whether a weapon is involved in their emergency—it is a vital piece of information for law enforcement to have as they roll up on the scene.

The collaborators strived to uncover “golden tweets,” which are defined as those that contain timely and actionable information tied to the 6Ws. The tool developed during the project automatically organizes tweets after they first have been sorted based on type of incident and location. After the initial sorting, the tweets are organized based on the 6 Ws, and then that information is displayed on a map. Ultimately, the hope is that tweet-organization tools like the one developed via this project can be integrated with an ECC’s computer-aided dispatch (CAD) system.

A key finding of the project was that telecommunicators and first responders in the field universally agreed that social media represents an increasingly important tool for augmenting emergency response, provided that the data is properly integrated and operationalized. Other key findings include the following:

• Seeing social media data integrated within an existing call-handling or dispatching platform is key. In Charleston, a RapidDeploy cloud-based CAD platform was used to demonstrate how data could be displayed in a CAD system user interface.
• The integration of new data, including social media, into ECC operations will require new, advanced training for multiple roles, including call-takers, dispatchers, supervisors, managers, and technical staff.
• While first responders in the field and telecommunicators were enthused about the opportunity to leverage social media, they offered varying opinions regarding the types of information they found most valuable given their specific emergency response role.
• New roles and skills may be necessary for monitoring, analyzing and operationalizing social media data. For example, an ECC might want to add a communications specialist role, which could provide career path progression for call-takers or dispatchers with strong analytical and problem-solving skills.

The much bigger challenge

Social media represents the proverbial tip of the iceberg concerning new types of data that will be available to the emergency response community, driven by broadband communications technology. Broadband will be the driving force behind systems that would choke a narrowband communications system, but which will generate mission-critical data in the future that will help emergency responders do their jobs better—resulting in more lives and property saved—and keep them safer.

To illustrate, let’s consider a hypothetical child abduction. The following are some of the systems that potentially could come into play, in addition to social media, all of which would enhance situational awareness dramatically:

• Automatic license plate reader system
• Camera-equipped unmanned aerial vehicles (i.e., drones)
• Geofencing systems
• Fixed, mobile and body-worn camera systems
• Facial-recognition systems (contingent on privacy laws)
• Vehicle-tracking systems
• AMBER alert systems, including integration with highway display signage

All of these systems and devices are available today. In the future, they will be supplemented by systems and devices that have yet to be tapped, notably Internet of Things (IoT) devices and systems, e.g., wearable medical devices and business/residential alarm systems.

This is going to require a sea change in thinking amongst public safety officials, as well as the government officials who control the amount of capital and operating expense money that flows into police departments, fire/rescue departments, emergency medical services agencies and ECCs to support communications technology evolution.

When it first became clear that public safety broadband communications were a real possibility, video was considered to be the “killer” app. How wonderful it would be, the conventional wisdom went, to have video from a plethora of sources—including citizen smartphones—pouring into an ECC or emergency operations center.

But that thinking has changed dramatically, driven by the notion that a torrent of unfiltered data streaming into an ECC and/or EOC would overwhelm telecommunicators and emergency responders in the field—it would be akin to drinking from the proverbial firehose.

Now the idea is to triage and analyze the data and then present it in a meaningful way into the ECC’s call workflow—as identified above, this too was a key goal of the pilot project—so that emergency responders can act effectively in real time. If the data cannot be harnessed effectively, then it has little value. But very sophisticated data-analytics solutions are needed to do this, and they still are at a nascent state, which is a problem.

Right now, MCP is working with the alarm-monitoring company, ADT, on a project that has a long-term goal of addressing the amount of data that will stream into ECCs and EOCs, which promises to increase exponentially, largely driven by sensors, which today number in the billions. In its research, ADT discovered that ECCs don’t want to comb through a tsunami of raw data to make decisions regarding who to send and where. What they want instead is a much smaller volume of highly contextual data, particularly video.

That context ideally would be delivered by improved data-analytics solutions, with actionable information displayed on a dashboard on a telecommunicator’s screen.  The dashboard information would be quick and easy to discern, which in turn will enable faster, easier and better decisions to be made concerning the type of emergency response to dispatch.

Dave Sehnert, ENP, is MCP’s director of innovation and integration. He can be emailed at DaveSehnert@MissionCriticalPartners.com.