Space Wombats| Spot That Fire V2.0

Project Details

The Challenge | Spot That Fire V2.0

Your challenge is to create an application that leverages NASA's near-real-time and archival wildfire datasets along with other tools to support firefighting and fire mitigation efforts. This challenge builds on last year’s challenge of the same name by calling for innovative ideas and apps that focus on how to engage and enable citizens to assist with the entire firefighting and fire mitigation process.

Fire Action Plan

Alexa app that uses NASA wildfire data to warn users and to provide a customisable fire action plan.

Space Wombats

In this project we used NASA's data on active wildfires and built a prototype of an Alexa app that monitors wildfires in the location of the device and capable of alerting the user whenever danger is present. There are 3 alert levels: low risk, medium risk and high risk. Low risk means there are no active wildfires around, medium risk means there is an active wildfire in a certain distance away and the user may choose to activate a fire action plan that would list required steps. High risk alert means there is an immediate danger and the user must leave immediately.

The actual distance to determine the danger levels were chosen arbitrarily and need to be reviewed with a Fire Fighting department.

There were multiple challenges faced during this hackathon:

1. Alexa was chosen as a mean of interaction with the user. None of the participants had experience with Alexa before.

2. Had to find a way to convert the user's address to geo coordinates and calculate a distance between the wildfire and the user.

3. Create a flow of a notification that would be as close to natural speech as possible.

The technology stack and flow is as follows:

1. AWS Cloud Watch triggers an event every 10 minutes that calls a poller Lambda function

2. The poller communicates to NASA's wildfire API and retrieves latest data

3. The data gets analysed by the lambda and verifies against the DynamoDB database users need to be notified.

4. If so, lambda puts a record in the database to alert user.

5. User can request the danger status via Alexa skill.

For further improvement and direction

It was an interesting project for us to work and we did discuss a lot of ideas to expand the project upon, here are some examples:

- Integrate the app with other providers of wildfire data

- Implement a modeling tool to predict wildfire progression

- Implement proactive notifications

- Turn the app into a social platform to assist with fire plan mitigations for a large group of people


The videos of the prototype can be found here:

https://www.youtube.com/watch?v=dlFADPnPrWw&feature=youtu.be


The source code can be found here:

https://github.com/AnthonyGDoueihi/fire_plan_api_parser

https://github.com/AnthonyGDoueihi/fire_plan