Impact of Storm Events on Morbidity, Mortality, and Economy

Data Processing

• Preprocessing

library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

• Download, Load, and Read Data The raw data is downloaded directly from the National Weather Service Website. The data contains events recorded from 1950 to 2011. The bz2 compressed data, is openend using read.csv, which is capable of opening compressed bz2 files.

if (!file.exists("StormData.csv.bz2")){
        data_url <- "https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2" 
        download.file(data_url, "StormData.csv.bz2")
}
StormData <- read.csv("StormData.csv.bz2")

• Health Impact Initially the morbidity and mortality rates were isolated and observed from the complete data set. These were filtered to eliminate those that had a zero occurrences to reduce the number of redundancies in the data. There were still a number of redundant event types: these were ignored in this program. The data for the number of Injuries and number of Deaths were plotted.

Event <- StormData %>% group_by(EVTYPE) %>% summarise(injuries = sum(INJURIES), deaths = sum(FATALITIES))
Event <- filter(Event, injuries>0 | deaths>0)
EvType <- filter(Event, deaths == max(Event$deaths))

• Economic Impact In observing the Economic impact, the values in the PROPDMG column need to be multiplied by the appropriate exponent to account for thousands, millions, or billions of dollars of damage. To determine the event that caused the most damage, the data were filtered similar to how they were filtered with the Mortality and Morbidity.

Econ <- StormData %>% group_by(EVTYPE, PROPDMG, PROPDMGEXP) %>% summarise(damage = sum(PROPDMG))
Econ <- filter(Econ, damage>0)
Econ <- if(Econ$PROPDMGEXP == "K") {
        mutate(Econ, damage = damage*10^3)
        
} else if(Econ$PROPDMGEXP == "M") {
        mutate(Econ, damage = damage*10^6)
} else if(Econ$PROPDMGEXP == "B") {
        mutate(Econ, damage = damage*10^9)
}

## Warning in if (Econ$PROPDMGEXP == "K") {: the condition has length > 1 and
## only the first element will be used

EcType <- filter(Econ, damage == max(Econ$damage))

Results

The morbidity and mortality of the different storm types are presented in the two figures. It is clear upon looking at them that there is one particular storm datum that stands out above the rest, and this is the same storm type in both the morbidity and mortality bar plots. Upon further investigation, the maximum number of injuries was found to be 9.134610^{4}, and the maximum number of deaths, 5633; both much greater than any other Event Type, and were a result of TORNADO.

The Disaster Event with the maximum economic impact was TORNADO, with an impact of 1.66110^{9} dollars.

par(mfrow = c(1,2))
with(Event, barplot(Event$injuries, xlab = "Storms by Type (unnamed)", ylab = "Number of Injuries"))
with(Event, barplot(Event$deaths, xlab = "Storms by Type (unnamed)", ylab = "Number of Deaths"))