• In 1994, the Northridge earthquake caused $75B and killed more than 60 people.
  • In 2017, Hurricane Irma caused $50B in damage to Miami and killed 129 people.
  • SDG Target 11.5 looks to significantly reduce the number of deaths and the number of people affected and substantially decrease the direct economic losses.
  • Disasters often impact communications infrastructure.

What it does

  • Disaster early warning.
  • Connect people with people and authority with people during a disaster via multiple communication channels.
  • Give suggestions to people and city authorities during the disaster.
  • Solutions for pre - during - after the disaster.

How we built it

  • By talking with Jason and other SDGs experts we all agree that communication during a disaster is a key component to reduce deads and losses. So we sat down and brainstorm about how technology can help.
  • After 2 hours the solution comes to light and with years of programming @iphong - a ReactJS developer he quickly built up the frontend.
  • @max80713 is the one who creates the most of our chatbot which user will get to talk during the disasters using Messenger APIs.
  • @Ngoc Khuat and @Hoàng Tùng Lâm (Linus) with the help of PyTorch experts built up the recommendation systems and the backend in a very short time.

Challenges we ran into

  • Figure out how to connect people together during the disaster.
  • Making the app accessible for all user.


  • Reduce cost and losses by early alert the disasters.
  • Connect people so they can help each other during the disaster.
  • Give city authorities a way to communicate with people during the disaster.

What we’ve learned

  • The damage of disaster on people lives and economic losses.
  • The importance of communication during disasters.
  • How great teamwork can make a big difference.

What's next for BeSafe

  • Expand it to other channels: SMS, Whatsapp, Mesh-Network.. so people can access it anywhere, anytime during the disaster.
  • Make the platform be easy to access, even with people in a less developed country.
  • Optimize to make the chatbot can respond quickly with millions of people in real time.
  • Collecting more data to improve the recommendation engine.
  • Using Wit.AI for better conversations.

Built With

+ 1 more
Share this project: