Where’s the human perspective in disaster relief?
These developments are inspiring and encouraging. But there are challenges, too. When mapping tools rely on tech alone, they help find routes to affected areas but miss the detailed –and changing– human perspective of what’s needed where and why. When they’re crowdsourced, they rely on the public to be informed and tech-savvy enough to know which tools, systems or apps to use in communicating information.
Citibeats wants to change this. We analyse citizens’ responses to disasters on the social media platforms they ordinarily use for immediate disaster response monitoring. We did a trial run in 2017, teaming up with NTT Data to interpret social media related to natural disasters across three cities in northern Japan: Akita, Aomori and Iwate. Our machine-learning algorithms followed the online conversation and instantly interpreted the data, to show which issues were of most concern.
We found that each city had different priorities. While in Aomori and Iwate access to healthcare was the primary concern, in Aikita, food and mobility were what was needed. Knowing where to focus resources first allows aid to be truly effective, targeted to the needs of the community.
In the weeks following a disaster, the city’s needs evolve and change. Citibeats artificial intelligence monitors the changing conversation, helping to rebuild the community long-term, as some concerns are resolved and new ones emerge. In Aikita, for instance, citizens first needed access to mobility, but in the later aftermath, they began to indicate the damage done to their houses and farmlands that needed fixing.
By hearing what citizens say, city leaders, insurance providers and aid organizations can make better decisions to reduce disaster impact in the short and long term. And looking even further forwards, this detailed record of how disasters impact citizens can even help to predict and prevent disaster damage in the future.