Synopsis

This analysis is part of Coursera’s Reproducible Research. Quoting the assignment,

“This project involves exploring the U.S. National Oceanic and Atmospheric Administration’s (NOAA) storm database. This database tracks characteristics of major storms and weather events in the United States, including when and where they occur, as well as estimates of any fatalities, injuries, and property damage.” The specific questions to be answered is what types of event most greatly effect population health, and which events most greatly effect the economy in the U.S.

The data is available here

Documentation and FAQ are available:

Data Processing

Load necessary packages

library(pacman)
p_load(dplyr, tidyr, stringr, lubridate, httr, ggvis, ggplot2, shiny, rio, rmarkdown)

Load the Data

fileURL <- "https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2"
download.file(fileURL, 
              destfile = "/Users/danieltheng/Desktop/Learning R/Reproducible Research/RepResearch-Course-project-2/StormData.csv.bz4")
StormData <- read.csv("/Users/danieltheng/Desktop/Learning R/Reproducible Research/RepResearch-Course-project-2/StormData.csv.bz4")

Which Event poses the greatest threat to population health?

Organizing the Data

Aggregate the Sum of Fatalities and Injuries by Event Type

fatal.by.evtype <- aggregate(FATALITIES ~ EVTYPE, data = StormData, sum)
injure.by.evtype <- aggregate(INJURIES ~ EVTYPE, data = StormData, sum)

Remove the Events with 0 Fatalities and Injuries

damage.by.evtype <- cbind(fatal.by.evtype, INJURIES = injure.by.evtype$INJURIES)
empty.ev <- which((damage.by.evtype$FATALITIES & damage.by.evtype$INJURIES) != 0)
damage.by.evtype1 <- damage.by.evtype[empty.ev, ]

Arrange Events by Fatalities, select top ten, gather Fatalities&Injuries

damage.by.evtype1 <- arrange(damage.by.evtype1, desc(FATALITIES))
damage.by.evtype1 <- damage.by.evtype1[1:10, ]
damage.by.evtype1 <- gather(damage.by.evtype1, CType, Value, c(FATALITIES, INJURIES))

Plot Bar Graph of the top 10 Event Types

g <- ggplot(damage.by.evtype1, aes(x = reorder(EVTYPE, -Value), 
                                   y = Value, 
                                   fill = (CType)))
g + geom_col() + labs(title = "Top Ten Events by Fatalities", 
                      x = "Type of Event",
                      y = "Counts",
                      fill = "Type") + 
    theme(axis.text.x = element_text(angle = 90, hjust = 1))

As the figure shows, Tornados cause the most injury and deaths in the United States. # Which Event has the great economic impact? ## Organizing the Data The data has the damage number with a code B, M, K for Billions, Millions, and Thousands respectively. We need to convert those damage codes into numerics.

econdmg <- StormData[ , c(8, 25:28)]
rm.empty <- which((econdmg$PROPDMG & econdmg$CROPDMG) != 0)
econdmg <- econdmg[rm.empty, ]
row.names(econdmg) <- 1:16242
K.prop <- grep("[Kk]", econdmg$PROPDMGEXP)
M.prop <- grep("[Mm]", econdmg$PROPDMGEXP)
B.prop <- grep("[Bb]", econdmg$PROPDMGEXP)
K.crop <- grep("[Kk]", econdmg$CROPDMGEXP)
M.crop <- grep("[Mm]", econdmg$CROPDMGEXP)
B.crop <- grep("[Bb]", econdmg$CROPDMGEXP)
econdmg[K.prop, 2] <- econdmg[K.prop, 2] * 1e4
econdmg[M.prop, 2] <- econdmg[M.prop, 2] * 1e6
econdmg[B.prop, 2] <- econdmg[B.prop, 2] * 1e12
econdmg[K.crop, 4] <- econdmg[K.crop, 4] * 1e4
econdmg[M.crop, 4] <- econdmg[M.crop, 4] * 1e6
econdmg[B.crop, 4] <- econdmg[B.crop, 4] * 1e12
econdmg <- econdmg[ , c(1, 2, 4)]

Aggregate

prop.dmg.evtype <- aggregate(PROPDMG ~ EVTYPE, data = econdmg, sum)
crop.dmg.evtype <- aggregate(CROPDMG ~ EVTYPE, data = econdmg, sum)
dmg.evtype <- cbind(prop.dmg.evtype, CROPDMG = crop.dmg.evtype$CROPDMG)

Arrange

dmg.evtype <- arrange(dmg.evtype, desc(PROPDMG))
dmg.evtype <- dmg.evtype[1:10, ]
dmg.evtype <- gather(dmg.evtype, Type, Value, c(PROPDMG, CROPDMG))

Plot

g <- ggplot(data = dmg.evtype, 
            aes(x = reorder(EVTYPE, -Value), 
                y = Value, 
                fill = (Type)))
g + geom_col()+ labs(title = "Top Ten Events by Economic Impact", 
                      x = "Type of Event",
                      y = "Damage in Dollars",
                      fill = "Type") + 
    theme(axis.text.x = element_text(angle = 90, hjust = 1))

As the figure shows Flood caused the most economic damage than another other event.

Results

Tornadoes cause the most injury and deaths in the U.S.
Floods cause the most economic damage in the U.S.

Analysis on the Health and Economic Impact of Weather Events in the U.S. with the NOAA Storm Database

dantheng

8/8/2020