Preparedness

Start-up Offers Artificial Intelligence to Help Emergency Responders During Disaster

Locating vulnerable populations and their vulnerabilities is what drives Geospiza.

by Jim McKay / August 31, 2018

Geospiza is one of a few startups who see a void in alerting and messaging in the emergency management realm and offer a solution in artificial intelligence software.

Sarah Tuneberg started Geospiza a little more than a year ago after years of acting as a consultant in the field of public safety and emergency management. During those years, she developed data models for determining who and where vulnerable populations are and what they are vulnerable to, and then developing solutions.

She started getting requests from all over for these models and decided it was a niche she needed to fill.

“It’s multi-sourced data integration to understand, in a more robust way on a geospatial level, who is especially vulnerable to disasters, why and what disasters they are vulnerable to and where they are located,” Tuneberg said. “The next place was what can we do to mitigate or prepare or reduce their vulnerability in some capacity.”

It’s a Web-based application and map interface that gleans data from multiple sources on vulnerable populations around 15 different hazard scenarios. The software maps the vulnerable populations in clusters considering factors such as non-English-speaking members of the community who are in an area vulnerable to flood.

The data is collected through a multitude of sources, including static sources such as the Census to much more live data like social media. In between, sources like the U.S. Department of Education for children and the Centers for Medicare and Medicaid Services to get information on people over 65.

“When we partner with a jurisdiction, we use their more robust geographically specific data,” Tuneberg said. Once the data is gleaned, it is mapped and the clusters emerge. “You see various geographies highlighted and various vulnerabilities and there are varying levels of granularity, depending on what data we can get,” she said.

When you click on the geography of interest, the software provides perspectives on things like transportation, communications and health-care matters as they relate to relative hazards — “such as where first responders can intervene and help people stay alive, who has a transportation vulnerability in this area and what are the leading contributors,” Tuneberg said.

That information is presented in a series of tables and maps along with a series of action recommendations.

Geospiza is working with agencies in Redmond, Wash., Multnomah County, Ore., and Jefferson County, Fla.

In Redmond, there is a large Vietnamese community, many of whom are non-English-speaking. Finding those in communities with such a vulnerability is one task, another is finding ways to communicate to them once they’re located.

“Just knowing who doesn’t speak English isn’t super helpful,” Tuneberg said. “What you then want to get into is when conducting extreme heat planning, for example, who, in addition to non-English-speaking residents, is at high vulnerability to extreme heat illness and where do they live.”

Geospiza is equipped to add value to large cities as well as smaller jurisdictions, Tuneberg said. For instance, one model involves a cyberattack that generates a large-scale power outage.
It’s an area, she says, where emergency management and public safety needs some help. “We’re not doing any better at this.”

  The 15 Hazard Scenarios:

•All hazards
•Complex Coordinated Attack
•Severe Summer Storm
•Extreme Heat
•Flood
•Hazardous Materials
•Hurricane
•IED
•Infrastructure Collapse
•Multiple Structure Fire (Urban Conflagration)
•Pandemic Influenza
•Power Outage
•Transportation Incident
•Utility Failure
•Wildland Fire
•Winter