Understanding and Predicting High-Risk Wildfire Zones

Inspiration

We were inspired by personal connections to the disasters in Maui to see if we could predict high-risk zones and prevent catastrophes in the future.

What it does

We present various visualizations and a risk-assessment model using gradient boosting to understand and predict wildfires in the United States from 2000-2010.

How we built it

We used Python, Pandas, and NumPy to clean our data, and we used Sci-Kit Learn to build our model, and Tableau to create our visualizations.

Challenges we ran into

We ran into a wall when trying to create visualizations using Tableau, which required our group to decide the best way to represent the data we wanted to show and which information we wanted to include. Joining a more precise geospatial table to our supplementary dataset introduced challenges as well for our team due to the difficulties that arose with trying to map out each data point.

Accomplishments that we're proud of

We successfully put together a machine learning model to determine which risk factors are most influential in wildfire development, as well as used Tableau for the first time to create stunning comprehensive visualizations.

What we learned

Many factors can influence the build-up to wildfires. By analyzing historical data on past wildfires and weather conditions, our group studied the primary causes and spread of these wildfires. We learned how machine learning models can be trained to recognize patterns used to identify areas prone to wildfires.

What's next for Understanding and Predicting High-Risk Wildfire Zones

Gathering global climate data to use in our model to better predict factors affecting wildfires. We want to contact high risk local and state governments in the future and create a web application to make our model accessible for any conditions and also predict wildfire risk for various localities daily.