AI and Social Media Enabling Rapid Disaster Response!

Disaster Response and Emergency Planning today needs the well-rounded support of AI and Social Media to function effectively. From developing situational awareness, early and continually, to bridging the gap between communication and infrastructure failures and just too much information, and even coordinating the relief efforts to those most prone, AI can literally help save humankind from all kinds of disasters.

Why Is This Important?

Natural disasters wreak havoc on human life and property every year. Some events cause billions of dollars’ worth of damage, mostly irrevocable. For example, the Northern California fires, the Hurricanes Irma and Harvey, and more recently, the volcanic eruption in Hawaii to name a few.

There are some typical areas of Disaster Management that Governments and agencies are constantly struggling with.

(a) Disaster Prevention: This requires that governments and agencies have their eyes and ears open, read through data collected from satellites, sensors, seismographs, drones or robots, and take informed action, (like directing evacuations or quarantine) quickly in order to contain the outcome.

(b) Disaster Identification: This requires that governments and agencies scour through heaps of data, keeping abreast of news updates, communicating with sources on ground zero, and identifying areas where life-saving response strategies have the best chance.

(c) Disaster Management: This requires coordinating relief work, sending aid, zoning the affected areas and communicating with the public, minimizing damage, enhancing the reach of rescue workers and making provisions for housing, hospital, safety and support.

The basis of effective Disaster Prevention, Identification and Management is solid data, but data is often highly fragmented, inaccessible or incomplete. It is either buried in too deep, scattered or available in varied formats. This makes the collation and analysis of data difficult by humans. But AI technology, on the other hand, can make navigation easy, viable and seamless.

The Role of Social Media

The role of Social Media in Disaster Response and Management is huge! It has been ascertained and confirmed that some of the most-actionable information during a crisis comes from social media users, who are acting as citizen journalists reporting the issue to their feed, or on-ground workers helping solve problems.

Twitter, Facebook, Instagram, and YouTube can quickly pick-up developing stories either through the hashtags people are using or by analysing images/videos and other data snippets that people are posting online. Digital content from these channels can flag, raise alarms, provide early warnings, identify exact location, and provide real-time reports of the event. For agencies, these channels can be leveraged for crowdsourcing relevant data.

AI and the Evidence-based Approach to Disaster Response

Data is abundant and pouring — whether from Social Media or from tens and hundreds of other channels that agencies employ. The need is to turn this data drawn from messy, real-world datasets into reliable information and actionable insights.

Building Heat-Maps: Machine Learning algorithms aggregate and consolidate the data, reconcile variable responses, and integrate and stitch together information derived from all channels to create a cohesive story. The analysis generates ‘heat maps’ of areas with heightened or abnormal activity. Qatar Computing Research Institute and 1 Concern are pioneering efforts in this area.

Image Processing and Deep Learning: Advanced Image Processing and Deep Learning techniques can analyse images for cognitive content such as — complex emotion in an image picked-up from Social Media that can help determine the extent of trauma experienced by people; and reading fault lines or changes in sea levels from satellite images that can help forecast the occurrence of a disaster.

ML and Natural Language Processing: By combining ML with NLP, further enhancements can be made available. These systems are highly intelligent and responsive. They enable users to ask specific questions pertaining to the event and receive useful answers. BlueLine Grid is a communication platform developed to assist rescue efforts that uses and leverages Natural Language Processing capabilities.

ML and Analytics: Analytics tools too can be highly useful. When combined with ML, Analytics can dramatically improve coordination and communication, opening up channels of information relay and data analysis on the go. WorkFusion’s coupled offering of ML and Analytics is delivering superior results in this area.