In 2010, when Haiti was rocked by an earthquake that killed over 150,000 people, aid workers manned SMS help lines where victims could reach out for help. Even with the international humanitarian effort, there was not enough manpower to effectively handle the volume of communication. We set out to fix that.
What it does
EmergAlert takes the place of a humanitarian volunteer at the phone lines, automating basic contact. It allows victims to request help, tell their location, place calls and messages to other people, and inform aid workers about their situation.
How we built it
We used Mix.NLU to create a Natural Language Understanding model that categorizes and interprets text messages, paired with the Smooch API to handle SMS and Slack contact. We use FHIR to search for an individual's medical history to give more accurate advice.
Challenges we ran into
Mentoring first time hackers was both a challenge and a joy.
Accomplishments that we're proud of
Coming to Canada.
What we learned
Project management is integral to a good hacking experience, as is realistic goal-setting.
What's next for EmergAlert
Bringing more depth to the NLU responses and available actions would improve the app's helpfulness in disaster situations, and is a good next step for our group.