The Textron and JP Morgan challenges inspired us to create an impactful disaster relief solution utilizing geospatial data.
What it does
Our program starts with a chatbot for user-interface that retrieves disaster victim data and sends it to google firebase which hosts our database. Victim data includes name, geo-coordinates, status, emergency contact information, and an informative message for relief organizations. Victims could also request forecasts of future conditions in weather. Government agencies or non-profit agencies are capable of retrieving victim data then input the number of relief camps they are capable of setting up and the program outputs the most optimal coordinates to place such camps.
How we built it
Our user interface relies on google diagflow to build the chatbot. Our database relies on google firebase to host our data. Our backend web app runs on flask. The machine learning clustering algorithm which determines best coordinates for relief camps utilizes sklearn. Our weather prediction utilizes accuweather's api. The dashboard for the web app is coded in react.js.
Challenges we ran into
From the start, we had no prior experience working with some apis and services. We had trouble retrieving data from google firebase. We had trouble determining the best way to cluster data points. Taking user inputs and predicting the weather inputs was also a challenge.
Accomplishments that we're proud of
The clustering algorithm worked as intended. It can find ideal spots for camps.
What we learned
We learned how to use certain apis like twilio, accuweather, and hosting apps, and analytics and prediction modeling.
What's next for Heads Up Disaster
Building into a proper application, such as a mobile app. We can use more data sets to make predictions better.