How Real-Time Information Can Assist Insurers During Natural Disasters

At every phase of disaster preparation and recovery, time is of the essence. Today, social media platforms allow authorities to share critical updates with the public instantaneously and coordinate relief efforts. It also allows those in need to request assistance from first responders. That’s what happened when Hurricane Harvey hit the United States in August 2017.

As the leading edge of the hurricane made landfall, local emergency systems became inundated with urgent requests for help. With no one aware of their plight, many people stopped making calls and started writing posts on social media. Nevertheless, according to research conducting by the Rice University Kinder Institute for Urban Research, when those affected by the storm switched from 911 to social media, first responders missed over 5,200 rescue requests, as they lacked the technology and procedures to gather, analyze and act on the posts.

In addition to missing thousands of posts from people in dire need, the Federal Emergency Management Agency’s (FEMA) damage estimates also failed to incorporate real-time data readily available on Twitter, leading to glaring differences between the government’s view of the disaster and the situation on the ground.

The Kinder Institute’s study shows that when governments and companies need an accurate, up-to-the-minute picture of local conditions, traditional data sources and systems designed before the advent of social media can struggle to keep pace with fast-moving events.

On the other hand, given the millions of smartphone users who turn to social media when the emergency infrastructure becomes overwhelmed, vast quantities of real-time data exist to help the public and private sector improve how they prepare and respond to significant incidents. Given the enormous amount of publicly available data, it’s becoming increasingly difficult for governments and corporations to justify not using social media alerts to guide their disaster recovery efforts, as well as inform the creation and design of the products and services they offer.

For the insurance industry, in particular, social media and publicly available data can play a more significant role in how the industry operates than it currently does today.

To deliver the products and services policyholders expect, the insurance industry must possess the ability to evaluate the risk associated with an event accurately and with confidence. More data leads to better catastrophe modeling, which supports more effective planning and disaster assistance, as well as the creation of prudent yet profitable policies.

Moreover, publicly available data shared via social media helps first responders, insurance companies, and government officials develop a granular view of a disaster’s impact, which in certain circumstances could help save lives and minimize damage to property.

The Kinder Institute’s study of the C.E. King Middle and High Schools is a case in point. According to the FEMA’s data, neither school appeared within the flood zone as a result of Hurricane Harvey. Yet pictures taken in the aftermath of the storm show severe flooding at both locations.

To understand why the FEMA data did not place the schools in the flood zone, researchers prepared maps of the area surrounding the schools, and then plotted flooding zones based on FEMA and Twitter data. Individuals who posted Tweets did not necessarily provide their addresses, which researchers considered as they created corresponding “buffer” zones around FEMA-plotted data and estimated locations of Tweets.

The maps incorporating Twitter data changed the damage estimates significantly to show that both schools were affected by flooding, as was actually the case. Furthermore, some zones showed flooding based on Twitter and FEMA data exclusively, while others showed an overlap, with both government data and social media posts confirming that flooding occurred in the same area.

As this case study shows, neither dataset replaces the other. However, by using both FEMA and Twitter data with the appropriate buffers applied, a far more accurate damage estimate emerges.

To be clear, data gathered from Twitter provides a way to corroborate as well as supplement FEMA data, not replace it. This is a particularly powerful concept if an insurance company publicizes the use of alternative data sources, such as social media alerts, as a means of streamlining the claims process and triggering timely policy payments.

In addition to creating a positive experience for existing policyholders, an insurance company’s willingness to embrace alternative data sources might entice customers in disaster-prone areas to purchase parametric insurance contracts, which pay out predetermined amounts when weather conditions exceed certain metrics, as would be the case in the event of a hurricane.

Ultimately, using social media alerts enables insurance companies to adopt a nimble, fact-based approach to disaster preparation and recovery that helps mitigate risk and deliver critical assistance to policyholders when they need it the most.

Ed Monan is director of Corporate Security Sales at Dataminr. He previously supported a variety of U.S. Government agencies, most recently the Department of Homeland Security.