We exploit a fundamental quality of humankind in order to accelerate recovery from natural disasters. Just one week ago, Hurricane Michael hit every State from Florida to Virginia, causing an estimated 29 deaths, many more people injured, and approximately 1200 people left without a home. How long will it take until these people will be able to go back home? And how much longer does it take in countries that are a lot poorer and less organised that the USA. Unfortunately, the government doesn't have infinite resources and, the despite all the efforts, officials take a very long time to bring everything back into place after a disaster. But we have a solution: cooperation.
What it does
Anonymous Heroes is a platform that provides a way to connect people willing to help to a city's needs after a disaster. On top of inviting all citizens to give a hand using an automatic emailing service that will email potential volunteers in the area, AH optimises their efforts to provide the best possible coverage of skills and services that a city and its people need to recover from a natural disaster.
How we built it
We are using Amazon Machine Learning to predict the impact of a catastrophe by parsing related news from the web. This prediction is then used to establish the number of people needed for physical damage and for medical damage (first aid) and split these people across different areas affected by the disaster. A React app for the frontend uses Cognito EAM to log the users into their accounts. Using the AWS Gateway API we then call a Lambda to retrieve the user's information from the DynamoDB. From an admin perspective, the web-scraping algorithm generates a file of data on the catastrophe which is then passed onto the DynamoDB. This will automatically trigger a call to a second Lambda to update the user side of the website if they live within reachable distance from the area of the disaster.
Challenges we ran into
Automatically triggering the Lambda when the DynamoDB table was updated. Synchronising table updates with user views. Getting a reasonable performance out of the Machine Learning algorithm.
Accomplishments that we're proud of
Our objective coming into this Hackathon was to have fun and challenge ourselves while trying to make something that could have a positive impact on the world. We think we have managed to do so :)
What we learned
How to use many of the AWS features that none of us was very familiar with.
What's next for Anonymous Heroes
There is thousands of improvement that could be implemented to make this service more effective for catastrophe recovery. Some of the ones that came to our mind include: efficient web scraping to improve the impact prediction performance, better UI, text message service (as an alternative to emails) for countries with little internet access, etc...