Albertamans

• 

  visualizations of wildfire spread, day 5

• 

  visualizations of wildfire spread, day 2

• 

  visualizations of wildfire spread, day 3

• 

  visualizations of wildfire spread, day 4

• 

  visualizations of wildfire spread, day 1

Inspiration

The Jasper fire in Alberta is harming thousnads of families and ruining the wildlife habitats there and our intention was to create something capable of raising awareness, and helping prevent these wildfires from getting out of control.

What it does

The AI takes in a request for a future day, and proceeds to analyze previous days of wildfire growth in Canada and makes a prediction as to the growth of the wildfire and creates a 2D model of it.

How we built it

We used sklearn and Tensorflow to train our model with the previous 30 days of data taken from the Canadian government to let the 2D Convolution make educated estimates as to the growth of the wildfires.

Challenges we ran into

We were initially planning on creating an app for spreading awareness, but neither of us had enough experience with app development to go through with it, so we ended up restarting halfway through and developed machine learning agents for the sake of modelling and prediction.

Accomplishments that we're proud of

Finishing the bare bones model with the limited time we have with high accuracy.

What we learned

A bit of Unity and Android Studio, and the introductory basics of Django. We also learnt more of matplotlib and how to be more specific with the plotting.

What's next for Wildfire Prediction AI

Continuing with our learning of machine learning, and possibly introduce reinforcement learning for a wider database, and convert it into an LLM for ease of use.

Built With

• python