Wildfires

Inspiration

Wanting to better forecast potential wildfires so preventive action can be taken.

What it does

This project uses R to model and forecast monthly wildfire occurrences based on historical data from 2000-2010. Time series methods like ARIMA are applied.

How we built it

The data was loaded into RStudio and aggregated to monthly totals. Packages like fabletools, forecast, and lubridate were used for time series analysis.

Challenges we ran into

Parsing the date fields required wrangling multiple date formats. Visualizing time series relationships was limited by data resolution. Errors arose when splitting data for training/testing

Accomplishments that we're proud of

Successfully aggregated data into monthly time series. Generated forecasts for 2010 using auto.arima() and assessed model accuracy. Practiced data wrangling skills like pivoting and filling missing dates.

What we learned

Important steps like checking data types, inspecting aggregates, and properly handling dates. Limits of a single time series for modeling. Benefits of modular code and iterative development.

What's next for Wildfires

Include other variables like location and fire causes. Apply models like Prophet and VAR. Use expanded dataset for model validation. Develop interactive visualizations of forecasts.

Built With

• r
• rstudio