As a response to the ongoing wildfires devastating vast areas of Australia, our team developed a web tool which provides wildfire data visualization, prediction, and logistics handling. We had two target audiences in mind: the general public and firefighters. The homepage of Phoenix is publicly accessible and anyone can learn information about the wildfires occurring globally, along with statistics regarding weather conditions, smoke levels, and safety warnings. We have a paid membership tier for firefighting organizations, where they have access to more in-depth information, such as wildfire spread prediction.

We deployed our web app using Microsoft Azure, and used Standard Library to incorporate Airtable, which enabled us to centralize the data we pulled from various sources. We also used it to create a notification system, where we send users a text whenever the air quality warrants action such as staying indoors or wearing a P2 mask.

We have taken many approaches to improving our platform’s scalability, as we anticipate spikes of traffic during wildfire events. Our codes scalability features include reusing connections to external resources whenever possible, using asynchronous programming, and processing API calls in batch. We used Azure’s functions in order to achieve this.

Azure Notebook and Cognitive Services were used to build various machine learning models using the information we collected from the NASA, EarthData, and VIIRS APIs. The neural network had a reasonable accuracy of 0.74, but did not generalize well to niche climates such as Siberia.

Our web-app was designed using React, Python, and d3js. We kept accessibility in mind by using a high-contrast navy-blue and white colour scheme paired with clearly legible, sans-serif fonts. Future work includes incorporating a text-to-speech feature to increase accessibility, and a color-blind mode. As this was a 24-hour hackathon, we ran into a time challenge and were unable to include this feature, however, we would hope to implement this in further stages of Phoenix.

Built With

Share this project: