Wildfire Analysis

Team

Hardik Mahipal Surana Pranshu Dave Tanvi Pisat

Installation

1. Setup a virtual environment

virtualenv -p python3 venv
source venv/bin/activate

2. Install project dependencies

pip3 install -r requirements.txt

3. Run regression models

python src/regression_models.py data/combined.csv

Dataset Features

Column Name	Description
FOD_ID	global unique identifier
FIRE_NAME	name of the incident
DISCOVERY_DATE	date when the fire was discovered/confirmed to exist
CONT_DATE	date when the fire was declared contained/controlled
STAT_CAUSE_DESCR	Description of the (statistical) cause of the fire
FIRE_SIZE	Estimate of acres within the final perimeter of the fire
FIRE_SIZE_CLASS	Code for fire size based on the number of acres burned
LATITUDE	in decimal degrees
LONGITUDE	in decimal degrees
STATE	Two-letter code where fire occured
distance	distance of sensors from fire
date_time	when the data was collected
air_temperature	air temperature (1 day reading)
relative_humidity	relative humidity (1 day reading)
wind_speed	wind speed (1 day reading)
precipitation	precipitation (1 day reading)

Other Features

Air Quality Air Quality Class/Category Air pressure

FFMC index DMC index DC index ISI index Buildup index

vegetation fuel

altitude/height above sea level

Potential Problems

1. Heteroscedasticity - When dependent variable's variability is not equal across values of an independent variable
2. Multicollinearity - When the independent variables are highly correlated to each other
3. Underfitting - high bias
4. Overfitting - high variance

Notes

1. date columns in the database are in julian days
2. FIRE_SIZE_CLASS - A=greater than 0 but less than or equal to 0.25 acres, B=0.26-9.9 acres, C=10.0-99.9 acres, D=100-299 acres, E=300 to 999 acres, F=1000 to 4999 acres, G=5000+ acres

Sources

1. Kaggle
2. AirNow API
3. Mesowest API