NOAA Storm Data Analysis

This data anaysis answers the following quesions.

• Across the United States, which types of events are most harmful with respect to population health?

• Across the United States, which types of events have the greatest economic consequences?

Synopsis

The top five storm types are mined from NOAA storm data with highest casualites. Casualties are either fatalites or injures. And the top five storm types are mined from NOAA storm data with highest economic damage. The economic damage is measured by property and crop damage.

Required Libraries

library(dplyr)
library(lattice)

Data Processing

Load Storm Data

Data is from the NOAA Storm Database starting in 1950 to 2011.

stormdata <- read.csv(bzfile("repdata-data-StormData.csv.bz2", open = "r"))
# wrap data frame into a data table
stormdata <- tbl_df(stormdata)
names(stormdata) <- tolower(names(stormdata))

Clean Up Economic Damage Values

The following assumptions were made for the propdmgexp and cropdmgexp variables.

Symbol	Exponent (10^x)
? or - or + or ? or 0	0
1	1
2 or h or H	2
3 or k or K	3
4	4
5	5
m or M or 6	6
7	7
m or M	8
b or B	9

# convert property damage exponents
stormdata$propdmgexp <- gsub("^$|-|\\+|\\?|0",  1E0, stormdata$propdmgexp)
stormdata$propdmgexp <- gsub("[1]",  1E1, stormdata$propdmgexp)
stormdata$propdmgexp <- gsub("[2|hH]", 1E2, stormdata$propdmgexp)
stormdata$propdmgexp <- gsub("[3|kK]", 1E3, stormdata$propdmgexp)
stormdata$propdmgexp <- gsub("[4]", 1E4, stormdata$propdmgexp) 
stormdata$propdmgexp <- gsub("[5]", 1E5, stormdata$propdmgexp) 
stormdata$propdmgexp <- gsub("[mM]|6", 1E6, stormdata$propdmgexp)
stormdata$propdmgexp <- gsub("[7]", 1E7, stormdata$propdmgexp)
stormdata$propdmgexp <- gsub("[8]", 1E8, stormdata$propdmgexp)
stormdata$propdmgexp <- gsub("[bB]", 1E9, stormdata$propdmgexp)

# apply exponents to damage estimates
stormdata$propdmg <- stormdata$propdmg * as.numeric(stormdata$propdmgexp)

# convert crop damage exponents
stormdata$cropdmgexp <- gsub("^$|-|\\+|\\?|0",  1E0, stormdata$cropdmgexp)
stormdata$cropdmgexp <- gsub("[1]",  1E1, stormdata$cropdmgexp)
stormdata$cropdmgexp <- gsub("[2|hH]", 1E2, stormdata$cropdmgexp)
stormdata$cropdmgexp <- gsub("[3|kK]", 1E3, stormdata$cropdmgexp)
stormdata$cropdmgexp <- gsub("[4]", 1E4, stormdata$cropdmgexp) 
stormdata$cropdmgexp <- gsub("[5]", 1E5, stormdata$cropdmgexp) 
stormdata$cropdmgexp <- gsub("[mM]|6", 1E6, stormdata$cropdmgexp)
stormdata$cropdmgexp <- gsub("[7]", 1E7, stormdata$cropdmgexp)
stormdata$cropdmgexp <- gsub("[8]", 1E8, stormdata$cropdmgexp)
stormdata$cropdmgexp <- gsub("[bB]", 1E9, stormdata$cropdmgexp)

# apply exponents to damage estimates
stormdata$cropdmg <- stormdata$cropdmg * as.numeric(stormdata$cropdmgexp)

Results

Total Casualites

A casualty is defined as a fatality or injury for this analysis.

total_casualties_by_evtype <- 
    stormdata %>% 
    group_by(evtype) %>% 
    summarise(total_casualties = sum(fatalities) + sum(injuries),
              total_fatalities = sum(fatalities), 
              total_injuries = sum(injuries)) %>%
    arrange(desc(total_casualties))
    
    # get top five events with the most casualties
    top5_casualties_by_evtype <- total_casualties_by_evtype[1:5,]

Plot Top Five Storm Events and the Human Fatalities and Injuries

barchart(data = top5_casualties_by_evtype, 
         total_fatalities + total_injuries ~ evtype, 
         stack=T, 
         xlab = "Storm Event Type", 
         ylab = "Total Casualites (fatalities + injuries)",
         main = "Top Five Human Casualties by Storm Event Type",
         auto.key = list(text=c("fatalities","injuries")))

The top five storm events that have been the most harmful to humans.

• Excessive Heat
• Flood
• Lighting
• Tornado
• TSTM Wind

Economic Impact of Storms

Total Economic Damage by Event Type

total_damage_by_evtype <- 
    stormdata %>% 
    group_by(evtype) %>% 
    summarise(total_dmg = sum(propdmg) + sum(cropdmg),
              total_prop_dmg = sum(propdmg), 
              total_crop_dmg = sum(cropdmg)) %>%
    arrange(desc(total_dmg))
    
    # get top five events with the most casualties
    total_damage_by_evtype <- total_damage_by_evtype[1:5,]

Plot Top 5 Storm Event Types that had Largest Econmoic Impact

barchart(data = total_damage_by_evtype, 
         total_prop_dmg + total_crop_dmg ~ evtype, 
         stack=T, 
         xlab = "Storm Event Type", 
         ylab = "Total Damage (property + crop)",
         main = "Top Five Storm Types with Greatest Economic Impact",
         auto.key = list(text=c("property damage","crop damage")))

The top five storm events that have had the greatest

• Flood
• Hail
• Hurricane/Typhoon
• Storm Surge
• Tornado