Join us for Code @ Think 2019 | San Francisco | February 12 – 15 Register now Limited availability
By IBM Developer Staff September 25, 2018
During the 7.1-magnitude earthquake in Mexico City last year, Subalekha Udayasakar experienced the helplessness and chaos that comes from being disconnected.
“So many people were offline and relief and other efforts couldn’t reach them,” she said. “Also there wasn’t any visibility into what was really going on with them.”
This observation inspired Subalekha and her teammates — Jonah Model, Katie Mathews, Gandharv Patil and Matthew Malin — to create a way to keep citizens and first-responders online during response and recovery.
Project Lantern, a two-part hardware and software solution, was named a top finalist in the 2018 Call for Code Global Challenge. The project relies on low-cost hardware, social data and real-time data that evolves as it learns and as situations change.
The lantern, a key-sized device, serves as a pop-up communication hub for disaster recovery over an offline, wireless network. With customizable web apps to receive news, ask for help, and volunteer, as well as a map tool to guide users to shelters, fresh water and fuel, the lantern is designed to keep communities organized in pivotal moments of need.
For the Call for Code challenge, the team used IBM Watson to make sense of the data collected offline and present it as a dashboard for first responders. IRIS, which stands for Intelligent Routing and Insights, aids collaboration between local volunteers, at risk populations and trained relief workers.
“What IRIS does is use all the data, as well as use the machine learning capabilities of IBM Watson and data from The Weather Channel and all the other APIs that are out there, to provide the relief worker with a conversation interface, a chatbot essentially, that they can use to understand what they do in that specific situation,” Subalekha said.
Under the hood is machine learning, powered by IBM Watson Studio, which tackles the problem of predicting the optimal distribution of supplies in affected areas. The solution pulls in Weather Company API data, and relies on a Cloudant database. IBM Watson AI Assistant powers the chatbot.
The IRIS dashboard displays real-time maps of disaster regions, travel routes and priorities across the long-range network. The chatbot helps find patterns and inform decisions each day about where to deliver supplies and how to serve those populations most in need.
The team is building a user experience that requires no formal training and can help organize a bottom-up (or top down) recovery effort. They are also working with trained experts in disaster response (one is even a team member) to ensure their solutions are at once pragmatic and reliable.
“To be part of a community like Call for Code is really exciting for us because it’s all these other people who are also interested in human progress and in building technologies that’s going to support them,” Subalekha said. “And if there’s anything that we can learn from each other, we want to be there, and we want to support that process.”
March 6, 2019
January 24, 2019
January 22, 2019
Back to top