Exploratory Data Analysis of Wildfires in USA
Abstract
Wildfires are one of the major natural/man-made disaster in the US, leading to economic losses and human suffering. Exploratory Data Analysis (EDA) refers to a process through which we extract knowledge from complex big data sets. The research work on National Fire Program Analysis (FPA) data is motivated by the increase in wildfires in USA. Analyzing wildfire activity in the United States (US) is a critical for a national cohesive wildfire planning. Estimating wildfire probability and distribution maps is the first step in wildfire risk assessment and mitigation. The research studies historical wildfire data sources to develop a data driven model to analyze the trend variability of wildfires in the United States. This research includes intensive data analysis to identify the causes of wildfires over years. The objective of this study is to compile and consolidate all the available historical wildfire trends, and to discover previously unknown patterns, relationships and insights. These findings may assist wildfire managers in better prevention strategies to help mitigate large wildfires by analyzing their causes.The analysis shows that annual number of wildfires has a stationary time series with minor fluctuations, but the number of major wildfires with acres greater than 5000 acres has been increasing rapidly nationally. A further time series analysis shows that 95\% of the wildfires happen between the April-October months of the year. The scope of this study includes data summary, descriptive statistics, temporal and spatial patterns of wildfires which has been investigated in the following sections.