Tornadoes Have Highest Human and Economic Impacts in the United States
Courtney D. Shelley
Reproducible Data: Peer Assessment #2
August 22, 2014
Synopsis
Storms and other severe weather events can result in personal injury and property damage. Predicting and preventing economic costs and injury is a vital mission of the U.S. National Oceanic and Atmospheric Administration (NOAA). Information regarding severe weather events can be inferred using the NOAA storm database, which 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. Data was collected from 1950-November 2011. Early years in the database contain fewer recorded events than in more recent years, though this is presumably due to a lack of good records rather than a significant increase in recent storm events.
Analysis included cumulative injury/fatality tolls and economic loss due to 15 of the most frequent storm events in the database (thunderstorms, tornados, heat, floods, cold, snow, hurricanes, hail, fire, fog, tidal events, “winter weather”, dust storms, rain, and drought).
Tornados were found to have the greatest impact both on humans and property/agriculture, causing 5633 deaths, 91,349 injuries, and over $3B in property and agricultural damage.
Data Processing
Data was obtained from the Coursera.org [Reproducible Data course website] [https://class.coursera.org/repdata-005/human_grading/view/courses/972590/assessments/4/submissions]. Data was downloaded on August 20, 2014 at 13:50 PST. Further documentation for the database is available at [NOAA Storm Events Database] [http://www.ncdc.noaa.gov/stormevents/details.jsp]. The dataset contains 902,297 observations of storm events. Only relevant datapoints were analyzed, including event type, fatalities, injuries, property damage, and crop damage. Event labels were standardized and the most frequently occuring event types were analyzed further. The clean data set was rewritten into a new .csv file.
library(data.table)
setwd("~/RepData_PeerAssessment2")
keyvars <- c("EVTYPE","FATALITIES","INJURIES","PROPDMG","PROPDMGEXP","CROPDMG","CROPDMGEXP")
myfile <- "repdata-data-StormData.csv"
stormy <- read.csv(myfile)[,keyvars]## Create consistent event type labeling
stormy$EVTYPE<-gsub("TSTM WIND", "THUNDERSTORM", stormy$EVTYPE) # THUNDERSTORM
stormy$EVTYPE<-gsub("THUNDERSTORM WINDS", "THUNDERSTORM", stormy$EVTYPE)
stormy$EVTYPE<-gsub("THUNDERSTORM WIND", "THUNDERSTORM", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MARINE THUNDERSTORM", "THUNDERSTORM", stormy$EVTYPE)
stormy$EVTYPE<-gsub("GUSTY THUNDERSTORM", "THUNDERSTORM", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SEVERE THUNDERSTORMS", "THUNDERSTORMS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("THUNDERSTORMS", "THUNDERSTORM", stormy$EVTYPE)
stormy$EVTYPE<-gsub("THUNDERSTORM/HAIL", "HAIL", stormy$EVTYPE) # HAIL
stormy$EVTYPE<-gsub("MARINE HAIL", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SLEET", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HAIL STORM", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SMALL HAIL", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/HAIL", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("NON SEVERE HAIL", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HAIL DAMAGE", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("THUNDERSTORM HAIL", "HAIL", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FLASH FLOOD", "FLOOD", stormy$EVTYPE) # FLOOD
stormy$EVTYPE<-gsub("FLOODING", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FLOOD/FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("URBAN/SML STREAM FLD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("URBAN/SMALL STREAM FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("URBAN FLOODS", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RIVER FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("COASTAL FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("LAKESHORE FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("URBAN FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FLOODS", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("DAM BREAK", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SMALL STREAM AND FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RURAL FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SMALL STREAM AND", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MAJOR FLOOD", "FLOOD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HIGH WINDS", "WINDS", stormy$EVTYPE) # WINDS
stormy$EVTYPE<-gsub("HIGH WIND", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Strong Wind", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("STRONG WINDS", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("STRONG WIND", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WIND", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WINDSS", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MARINE WINDS", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WINDS ADVISORY", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("GUSTY WINDS", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HIGH WINDS", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("GRADIENT WINDS", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("GUSTNADO", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WINDS DAMAGE", "WINDS", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAVY RAIN", "RAIN", stormy$EVTYPE) # RAIN
stormy$EVTYPE<-gsub("Heavy Rain", "RAIN", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAVY RAINS", "RAIN", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MONTHLY RAINFALL", "RAIN", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FREEZING RAIN", "RAIN", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MONTHLY PRECIPITATION", "RAIN", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MIXED PRECIPITATION", "RAIN", stormy$EVTYPE)
stormy$EVTYPE<-gsub("LIGHT RAIN", "RAIN", stormy$EVTYPE)
stormy$EVTYPE<-gsub("BLIZZARD", "SNOW", stormy$EVTYPE) # SNOW
stormy$EVTYPE<-gsub("HEAVY SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Heavy Snow", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MODERATE SNOWFALL", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("LAKE-EFFECT SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("LAKE EFFECT SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("LIGHT SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW SQUALLS", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("ICE/SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FIRST SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW SHOWERS", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("EARLY SNOWFALL", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("ICE ON ROAD", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("EXCESSIVE SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Snow", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Light Snow", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/COLD", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW SQUALL", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW-SQUALLS", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/SQUALLS", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/BLOWING", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RECORD SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/BLOWING SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW AND SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAVY LAKE SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/COLD", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/RAIN", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW RAIN", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Light SNOW", "SNOW", stormy$EVTYPE)
stormy$EVTYPE<-gsub("EXTREME COLD", "COLD", stormy$EVTYPE) # COLD
stormy$EVTYPE<-gsub("RECORD COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("COLD/WINDS CHILL", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FROST/FREEZE", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("ICE STORM", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("EXTREME WINDSCHILL", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("PROLONG COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("COLD WINDS CHILL TEMPERATURES", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("UNSEASONABLY COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("UNUSUALLY COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FREEZE", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WINDS CHILL", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FROST", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HARD COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("COLD WEATHER", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("BLACK ICE", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("ICE", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WINDS/COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("COLD TEMPERATURES", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("EXTREME COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SEVERE COLD", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("COLD WAVE", "COLD", stormy$EVTYPE)
stormy$EVTYPE<-gsub("EXTREME HEAT", "HEAT", stormy$EVTYPE) # HEAT
stormy$EVTYPE<-gsub("EXCESSIVE HEAT", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RECORD WARM", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RECORD WARMTH", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEATTH", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("UNSEASONABLY WARM", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("PROLONG WARMTH", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RECORD HEAT", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("VERY WARM", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAT WAVE", "HEAT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WILD/FOREST FIRE", "FIRE", stormy$EVTYPE) # FIRE
stormy$EVTYPE<-gsub("WILDFIRE", "FIRE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("DENSE SMOKE", "FIRE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SMOKE", "FIRE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FIRES", "FIRE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("DENSE FOG", "FOG", stormy$EVTYPE) # FOG
stormy$EVTYPE<-gsub("FREEZING FOG", "FOG", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WINTER WEATHER/MIX", "WINTER WEATHER", stormy$EVTYPE) # WINTER WEATHER
stormy$EVTYPE<-gsub("WINTRY MIX", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("WINTER WEATHER MIX", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Winter Weather", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SNOW/RAIN", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HAIL/RAIN", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RAIN/HAIL", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAVY MIX", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("MIXED PRECIP", "WINTER WEATHER", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HURRICANE/TYPHOON", "HURRICANE", stormy$EVTYPE) # HURRICANE
stormy$EVTYPE<-gsub("TROPICAL STORM", "HURRICANE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("TROPICAL DEPRESSION", "HURRICANE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("TYPHOON", "HURRICANE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("RIP CURRENT", "TIDE", stormy$EVTYPE) # TIDE
stormy$EVTYPE<-gsub("STORM SURGE/TIDE", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HIGH SURF", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("STORM SURGE", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("ASTRONOMICAL HIGH TIDE", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("ASTRONOMICAL LOW TIDE", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("SEICHE", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("TSUNAMI", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAVY SURF/TIDE", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("TIDES", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("TIDE ADVISORY", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HIGH SEAS", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("TIDE ADVISORIES", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAVY SURF", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HAZARDOUS SURF", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HIGH WATER", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("High Surf", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("COASTAL FLOODING", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Coastal Flooding", "TIDE", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FUNNEL CLOUDS", "TORNADO", stormy$EVTYPE) # TORNADO
stormy$EVTYPE<-gsub("TORNADO DEBRIS", "TORNADO", stormy$EVTYPE)
stormy$EVTYPE<-gsub("FUNNEL", "TORNADO", stormy$EVTYPE)
stormy$EVTYPE<-gsub("TORNADO CLOUD", "TORNADO", stormy$EVTYPE)
stormy$EVTYPE<-gsub("VOLCANIC ASHFALL", "VOLCANIC ASH", stormy$EVTYPE) # VOLCANIC ASH
stormy$EVTYPE<-gsub("DUST DEVIL", "DUST STORM", stormy$EVTYPE) # DUST STORM
stormy$EVTYPE<-gsub("BLOWING DUST", "DUST STORM", stormy$EVTYPE)
stormy$EVTYPE<-gsub("ABNORMALLY DRY", "DROUGHT", stormy$EVTYPE) # DROUGHT
stormy$EVTYPE<-gsub("DRY MICROBURST", "DROUGHT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("UNSEASONABLY DRY", "DROUGHT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("HEAT/DROUGHT", "DROUGHT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("DRY", "DROUGHT", stormy$EVTYPE)
stormy$EVTYPE<-gsub("Other", "OTHER", stormy$EVTYPE)
storm<-data.frame(stormy$EVTYPE, stormy[,3:8])
exp<-storm$PROPDMGEXP ## Merge PROPDMG and PROPDMGEXP
exp[is.na(exp)] = 1 # Replace NAs with 1 for later column multiplication
Exp <- as.numeric(exp) # Converts B to 2, K to 17, and M to 19
Exp[Exp == 17] = 1000 # Replace K with 1000
Exp[Exp == 19] = 1000000 # Replace M with 1000000
Exp[Exp == 2] = 1000000000 # Replace B with 1000000
PropDMG <- storm$PROPDMG
PropDMG[is.na(PropDMG)] = 0
storm$PROPDMG <- PropDMG * Exp
Exp2<-storm$CROPDMGEXP # Crop Damage
levels(Exp2) <- c(levels(Exp2), 1, 1000, 1000000, 1000000000)
Exp2[is.na(Exp2)] = 1 # Replace NAs with 1 for later column multiplication
Exp2 <- as.numeric(Exp2) # Converts K to 17 and M to 19
Exp2[Exp2 == 17] = 1000 # Replace K with 1000
Exp2[Exp2 == 19] = 1000000 # Replace M with 1000000
Exp2[Exp2 == 2] = 1000000000 # Replace B with 1000000000
CropDMG<- storm$CROPDMG
CropDMG[is.na(CropDMG)] = 0
storm$CROPDMG <- CropDMG * Exp2
stormData<-data.table(stormy$EVTYPE, stormy[,3:5], stormy[,7])
setnames(stormData, old=c("V1", "FATALITIES", "INJURIES", "PROPDMG", "V3"),
new=c("EVTYPE", "FATALITIES", "INJURIES", "PROPDMG", "CROPDMG"))
write.csv(stormData, "stormData.csv")
setkey(stormData, EVTYPE)ANALYSIS
The most frequent storm types were thunderstorms, hail, floods, rain, cold, heat, fire, fog, “winter weather”, hurricane, tidal events, tornados, volcanic ash, dust storms, and droughts. Total human and economic damages were calculated for each event. Thunder was omitted from further analysis because no human or economic damage was attributed to thunder events. Cumulative human and ecomonic impacts are presented in the following table, preceded by supporting R code.
mainEventTypes<-c("HAIL", "FLOOD", "RAIN", "SNOW", "COLD", "HEAT", "FIRE", "FOG",
"WINTER WEATHER", "HURRICANE", "TIDE", "TORNADO", "DUST STORM",
"DROUGHT")
EVTYPE<-character(); FATALITIES<-numeric(); INJURIES<-numeric();
PROPERTY<-numeric(); CROPDMG<-numeric()
for(i in mainEventTypes) {
s<-subset(stormData, stormData$EVTYPE==i)
FAT<-sum(s$FATALITIES)
INJ<-sum(s$INJURIES)
PROP<-sum(s$PROPDMG)
CROP<-sum(s$INJURIES)
EVTYPE<-append(EVTYPE, i, after=length(EVTYPE))
FATALITIES<-append(FATALITIES, FAT, after=length(FATALITIES))
INJURIES<-append(INJURIES, INJ, after=length(INJURIES))
PROPERTY<-append(PROPERTY, PROP, after=length(PROPERTY))
CROPDMG<-append(CROPDMG, CROP, after=length(CROP))
}
DT<-data.table(EVTYPE, FATALITIES, INJURIES, PROPERTY, CROPDMG)
DT## EVTYPE FATALITIES INJURIES PROPERTY CROPDMG
## 1: HAIL 22 1467 698416 1467
## 2: FLOOD 1547 8674 2454784 32
## 3: RAIN 105 274 55070 482
## 4: SNOW 237 1896 174446 91349
## 5: COLD 541 2463 89621 954
## 6: HEAT 3121 9156 3033 1666
## 7: FIRE 87 1456 124524 641
## 8: FOG 80 1076 17259 1076
## 9: WINTER WEATHER 64 641 17774 1456
## 10: HURRICANE 183 1666 71944 9156
## 11: TIDE 792 954 37043 2463
## 12: TORNADO 5633 91349 3212453 1896
## 13: DUST STORM 24 482 5770 274
## 14: DROUGHT 3 32 5837 8674Query 1: Storm Types Which Inflict the Most Harm to Population Health
###### Query 1: Which storm types do the most population damage? ######
query1<-DT[order(INJURIES)] # Order by damage
names<-c("DROUGHT", "RAIN", "DUST STORM", "WINTER WEATHER", "TIDE", "FOG", "FIRE", "HAIL",
"HURRICANE", "SNOW", "COLD", "FLOOD", "HEAT", "TORNADO")
par(mar=c(5,9,5,5))
barplot(query1$INJURIES, horiz=TRUE, col = c("darkslategray3"), names.arg=names, las=1, xlim=c(0,90000), xlab="HUMAN IMPACT", main="INJURIES AND FATALITIES DUE TO STORM TYPES")
par(new=TRUE)
barplot(query1$FATALITIES, horiz=TRUE, col = "red", xlim=c(0,90000))
legend("bottomright", pch=22, fill=c("darkslategray3", "red"), legend = c("Injuries", "Fatalities")) 
FIGURE 1: Injuries and Fatalities Due to Most Common Storm Events. Cumulative health tolls of most common storm events from 1950-November 2011. Tornados cause the most human devastation, both in number of injuries caused and in number of deaths. Heat and floods also have large impacts on human health.
Query 2: Storm Types Which Inflict the Most Economic Harm
###### Query 2: Economic Damage ######
query2<-DT[order(PROPERTY)] # Order by cost
names<-c("DROUGHT", "RAIN", "DUST STORM", "WINTER WEATHER", "TIDE", "FOG", "FIRE", "HAIL",
"HURRICANE", "SNOW", "COLD", "FLOOD", "HEAT", "TORNADO")
par(mar=c(5,9,5,5))
barplot(query2$PROPERTY, horiz=TRUE, col = c("gold2"), names.arg=names, las=1, xlim=c(0,3.5e6), xlab="ECONOMIC IMPACT in US$", main="PROPERTY AND CROP DAMAGE DUE TO STORM TYPES")
par(new=TRUE)
barplot(query2$CROP, horiz=TRUE, col = "darkorange4", xlim=c(0,3.5e6))
legend("bottomright", pch=22, fill=c("gold2", "darkorange4"), legend = c("Property", "Crops"))
FIGURE 2: Property and Agricultural Damage Due to Most Common Storm Events. Cumulative property and agricultural damage (measured in US dollars) of most common storm events from 1950-November 2011. Tornados cause the most property damage while cold (including freezes and frosts) had the greatest agricultural impact.
Results
Tornados were found to have the largest impact both in terms of human health and economic loss. Tornadoes caused the highest number of fatalities (5633 within the period 1950-November 2011), and the greatest number of injuries (91,349 injuries attributable to tornadoes within the study period). Tornadoes also caused the greatest loss of property ($3.2 billion).