Strom Event - Public Health & Economic impact analysis

SYNOPSIS

# Storms and other severe weather events can cause both public health and economic problems for communities and municipalities. Many severe events can result in fatalities, injuries, and property damage, and preventing such outcomes to the extent possible is a key concern.The analysis shows the impact of weather events were flash floods and thunderstorm winds caused billions of dollars in property damages between 1950 and 2011. The largest damage to crops were caused by droughts, followed by floods and hailing.

DATA PROCESSING

#Loading raw data

library(“ggplot2”) library(“R.utils”) library(reshape2) library(“knitr”) url <- “https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2” download.file(url, “StormData.csv.bz2”) bunzip2(“StormData.csv.bz2”, “StormData.csv”)

main_data <- read.csv(“StormData.csv”)

Dim (main_data) head (main_data)

colnames(main_data) <- tolower(colnames(main_data))


data <- subset(x=main_data, 
               subset=(evtype != "?" & 
                           (injuries > 0 | fatalities > 0 | propdmg > 0 | cropdmg > 0)),
               select=c("evtype", 
                        "fatalities", 
                        "injuries", 
                        "propdmg", 
                        "propdmgexp", 
                        "cropdmg", 
                        "cropdmgexp"))   

# Health Data aggregation

data\(propdmgexp <- toupper(data\)propdmgexp) data\(cropdmgexp <- toupper(data\)cropdmgexp)

# Map property damage alphanumeric exponents to numeric values.
propDmgKey <-  c("\"\"" = 10^0,
                 "-" = 10^0, 
                 "+" = 10^0,
                 "0" = 10^0,
                 "1" = 10^1,
                 "2" = 10^2,
                 "3" = 10^3,
                 "4" = 10^4,
                 "5" = 10^5,
                 "6" = 10^6,
                 "7" = 10^7,
                 "8" = 10^8,
                 "9" = 10^9,
                 "H" = 10^2,
                 "K" = 10^3,
                 "M" = 10^6,
                 "B" = 10^9)
data$propdmgexp <- propDmgKey[as.character(data$propdmgexp)]
data$propdmgexp[is.na(data$propdmgexp)] <- 10^0

# Map crop damage alphanumeric exponents to numeric values
cropDmgKey <-  c("\"\"" = 10^0,
                 "?" = 10^0, 
                 "0" = 10^0,
                 "K" = 10^3,
                 "M" = 10^6,
                 "B" = 10^9)
data$cropdmgexp <- cropDmgKey[as.character(data$cropdmgexp)]
data$cropdmgexp[is.na(data$cropdmgexp)] <- 10^0

# Aggregate number of fatalities and injuries per evtype into healthData dataframe healthData <- aggregate(cbind(fatalities, injuries) ~ evtype, data=data, FUN=sum) # Add total column to healthData healthData\(total <- healthData\)fatalities + healthData$injuries

# Remove rows with zero health impact healthData <- healthData[healthData$total > 0, ] # Sort health data in descending order healthData <- healthData[order(healthData$total, decreasing=TRUE), ] # Re-label the rows rownames(healthData) <- 1:nrow(healthData) # Create dataframe of highest health impacting event types and append an “other” event type as a catchall # for everything else healthDataTop <- healthData[1:10, ]

# Economic data aggregation
data$propertyloss <- data$propdmg * data$propdmgexp
data$croploss <- data$cropdmg * data$cropdmgexp
economicData <- aggregate(cbind(propertyloss, croploss) ~ evtype, data=data, FUN=sum)
economicData$total <- economicData$propertyloss + economicData$croploss

# Remove rows with zero economic impact economicData <- economicData[economicData$total > 0, ] # Sort the economy data in descending order economicData <- economicData[order(economicData$total, decreasing=TRUE), ] # Re-label the rows rownames(economicData) <- tolower(rownames(economicData)) # Create dataframe of highest economy impacting event types economicDataTop <- economicData[1:10, ]

RESULTS

#Ready to plot data

healthDataTopMelt <- melt(healthDataTop, id.vars=“evtype”)

# Plot

healthChart <- ggplot(healthDataTopMelt, aes(x=reorder(evtype, -value), y=value))

healthChart = healthChart + geom_bar(stat=“identity”, aes(fill=variable), position=“dodge”)

healthChart = healthChart + scale_y_sqrt(“Frequency Count”)

healthChart = healthChart + xlab(“Event Type”) healthChart = healthChart + theme(axis.text.x = element_text(angle=45, hjust=1)) healthChart = healthChart + ggtitle(“Pareto Chart of Top 10 US Storm Health Impacts”)

print (healthChart)

# Top 10 Economic impact

# Prepare plot data economicDataTopMelt <- melt(economicDataTop, id.vars=“evtype”)

economicChart <- ggplot(economicDataTopMelt, aes(x=reorder(evtype, -value), y=value))
# Add bars                            
economicChart <- economicChart + geom_bar(stat="identity", aes(fill=variable), position="dodge")
# Format y-axis scale and set y-axis label
economicChart <- economicChart + scale_y_sqrt("Damage Impact [$]") 
# Set x-axis label
economicChart <- economicChart + xlab("Event Type") 
# Rotate x-axis tick labels 
economicChart <- economicChart + theme(axis.text.x = element_text(angle=45, hjust=1))
# Set chart title
economicChart <- economicChart + ggtitle("Pareto Chart of Top 10 US Storm Economic Impacts")
# Display the chart
print(economicChart)

SUMMARY

#Tornado and flood are contributed in damaging human health interms of fatalites and injuries and floor directly affected the human life with huge economical impact.