Health and Economic Consequences of Weather Events
N.J. Blume
Tuesday, February 17, 2015
Synopsis
The effects of weather events on local health and property is analyzed using the U.S. National Oceanic and Atmospheric Administration’s (NOAA) storm database. Tornados, warm weather (i.e. heat), flood and high winds cause the largest number of fatalities and injuries across events. Heat and coastal storms are the deadliest per event. Hurricanes and heat are the most injurious per event. Floods, hurricanes, tornados and hazardous surf caused the most property damage. Excessively dry weather, floods, ice and hurricanes caused the most crop damage.
Environment
R version TK Patched (TK)
OS Microsoft Windows 8.1
Platform: Dell Inspiron 3521 BLUEICING (64-bit) Intel Pentium CPU 2127U@1.90GHz
Locale: en_US.UTF-8/ TK get this from command
attached base packages: stats graphics grDevices utils dfs base
other attached packages:
require(ggplot2)## Loading required package: ggplot2## Warning: package 'ggplot2' was built under R version 3.1.1loaded via a namespace (and not attached): NA
Version Control
local: Git Bash, ~/datasciencecoursera/ReproducibleResearch
remote: RPub.com/NJBlume
Data Processing
The data are imported below.
setwd("~/..")
setwd("datasciencecoursera/ReproducibleResearch")
d <- read.csv(file="repdata-data-StormData.csv", head=TRUE, sep=",", stringsAsFactors=FALSE)
df<-d[ , c("EVTYPE", "FATALITIES", "INJURIES", "PROPDMG", "PROPDMGEXP", "CROPDMG", "CROPDMGEXP")]
df = as.data.frame(sapply(df, tolower))EVTYPE describes the type of weather event that caused some type of damage. Misspellings and alternative terminology are fixed below. Some closely related events are also grouped.
df <- df[-grep("summary", df$EVTYPE), ]
df[grep("sleet", df$EVTYPE), "EVTYPE2"]<-"sleet"
df[grep("freezing", df$EVTYPE), "EVTYPE2"]<-"sleet"
df[grep("freeze", df$EVTYPE), "EVTYPE2"]<-"ice"
df[grep("frost", df$EVTYPE), "EVTYPE2"]<-"ice"
df[grep("ice", df$EVTYPE), "EVTYPE2"]<-"ice"
df[grep("icy", df$EVTYPE), "EVTYPE2"]<-"ice"
df[grep("avalance", df$EVTYPE), "EVTYPE2"]<-"avalanche"
df[grep("avalanche", df$EVTYPE), "EVTYPE2"]<-"avalanche"
df[grep("thunderstorm", df$EVTYPE), "EVTYPE2"]<-"thunderstorm"
df[grep("tstm", df$EVTYPE), "EVTYPE2"]<-"thunderstorm"
df[grep("dryness", df$EVTYPE), "EVTYPE2"]<-"excessively dry"
df[grep("dry", df$EVTYPE), "EVTYPE2"]<-"excessively dry"
df[grep("driest", df$EVTYPE), "EVTYPE2"]<-"excessively dry"
df[grep("low rainfall", df$EVTYPE), "EVTYPE2"]<-"excessively dry"
df[grep("drought", df$EVTYPE), "EVTYPE2"]<-"excessively dry"
df[grep("wind", df$EVTYPE), "EVTYPE2"]<-"wind"
df[grep("wnd", df$EVTYPE), "EVTYPE2"]<-"wind"
df[grep("snow", df$EVTYPE), "EVTYPE2"]<-"snow"
df[grep("flood", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("hot", df$EVTYPE), "EVTYPE2"]<-"warm weather"
df[grep("heat", df$EVTYPE), "EVTYPE2"]<-"warm weather"
df[grep("warmth", df$EVTYPE), "EVTYPE2"]<-"warm weather"
df[grep("warm", df$EVTYPE), "EVTYPE2"]<-"warm weather"
df[grep("high temp", df$EVTYPE), "EVTYPE2"]<-"warm weather"
df[df$EVTYPE == "high", "EVTYPE2"]<-"warm weather"
df[grep("record high", df$EVTYPE), "EVTYPE2"]<-"warm weather"
df[grep("heatburst", df$EVTYPE), "EVTYPE2"]<-"warm weather"
df[grep("cold", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("low temp", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("record low", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("winter weather", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("wintry", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("wintery", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("cool", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("winter mix", df$EVTYPE), "EVTYPE2"]<-"cold weather"
df[grep("precip", df$EVTYPE), "EVTYPE2"]<-"rain"
df[grep("rain", df$EVTYPE), "EVTYPE2"]<-"rain"
df[grep("wet", df$EVTYPE), "EVTYPE2"]<-"rain"
df[grep("shower", df$EVTYPE), "EVTYPE2"]<-"rain"
df[grep("hail", df$EVTYPE), "EVTYPE2"]<-"hail"
df[grep("coastal erosion", df$EVTYPE), "EVTYPE2"]<-"rock slide"
df[grep("fog", df$EVTYPE), "EVTYPE2"]<-"fog"
df[grep("vog", df$EVTYPE), "EVTYPE2"]<-"fog"
df[grep("surf", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("swell", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("storm surge", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("red", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("sea", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("rogue wave", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("seiche", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("tide", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("wave", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("current", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("coastal surge", df$EVTYPE), "EVTYPE2"]<-"hazardous surf"
df[grep("flood", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("fldg", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("fld", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("stream", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("floooding", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("high water", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("rising water", df$EVTYPE), "EVTYPE2"]<-"flood"
df[grep("funnel", df$EVTYPE), "EVTYPE2"]<-"wall cloud/funnel cloud"
df[grep("wall cloud", df$EVTYPE), "EVTYPE2"]<-"wall cloud/funnel cloud"
df[grep("hurricane", df$EVTYPE), "EVTYPE2"]<-"hurricane"
df[grep("floyd", df$EVTYPE), "EVTYPE2"]<-"hurricane"
df[grep("smoke", df$EVTYPE), "EVTYPE2"]<-"smoke"
df[grep("volcanic", df$EVTYPE), "EVTYPE2"]<-"volcanic eruption"
df[grep("tornado", df$EVTYPE), "EVTYPE2"]<-"tornado"
df[grep("torndao", df$EVTYPE), "EVTYPE2"]<-"tornado"
df[grep("lightning", df$EVTYPE), "EVTYPE2"]<-"lightning"
df[grep("lighting", df$EVTYPE), "EVTYPE2"]<-"lightning"
df[grep("lignting", df$EVTYPE), "EVTYPE2"]<-"lightning"
df[grep("ligntning", df$EVTYPE), "EVTYPE2"]<-"lightning"
df[grep(" lightning", df$EVTYPE), "EVTYPE2"]<-"lightning"
df[grep("fire", df$EVTYPE), "EVTYPE2"]<-"wildfire"
df[grep("dust", df$EVTYPE), "EVTYPE2"]<-"blowing dust"
df[grep("blizzard", df$EVTYPE), "EVTYPE2"]<-"blizzard"
df[grep("fire", df$EVTYPE), "EVTYPE2"]<-"wildfire"
df[grep("slide", df$EVTYPE), "EVTYPE2"]<-"mud/rock slide"
df[grep("erosion", df$EVTYPE), "EVTYPE2"]<-"mud/rock slide"
df[grep("erosin", df$EVTYPE), "EVTYPE2"]<-"mud/rock slide"
df[grep("landslump", df$EVTYPE), "EVTYPE2"]<-"mud/rock slide"
df[grep("spout", df$EVTYPE), "EVTYPE2"]<-"waterspout"
df[grep("tropical storm", df$EVTYPE), "EVTYPE2"]<-"tropical storm"
df[grep("dam", df$EVTYPE), "EVTYPE2"]<-"dam failure"
df[grep("no severe", df$EVTYPE), "EVTYPE2"]<-"mild pattern"
df[grep("none", df$EVTYPE), "EVTYPE2"]<-"mild pattern"
df[grep("urban/small", df$EVTYPE), "EVTYPE2"]<-"mild pattern"
df[grep("urban and small", df$EVTYPE), "EVTYPE2"]<-"mild pattern"
df[grep("urban small", df$EVTYPE), "EVTYPE2"]<-"mild pattern"
df[grep("gustnado", df$EVTYPE), "EVTYPE2"]<-"gustnado"
df[grep("severe turbulence", df$EVTYPE), "EVTYPE2"]<-"other"
df[grep("microburst", df$EVTYPE), "EVTYPE2"]<-"other"
df[grep("downburst", df$EVTYPE), "EVTYPE2"]<-"other"
df[grep("metro storm", df$EVTYPE), "EVTYPE2"]<-"other"
df[grep("northern lights", df$EVTYPE), "EVTYPE2"]<-"other"
df[grep("marine", df$EVTYPE), "EVTYPE2"]<-"?"
df[df$EVTYPE == "coastalstorm", "EVTYPE2"]<-"coastal storm"
df[df$EVTYPE == "excessive", "EVTYPE2"]<-"?"
df[df$EVTYPE == "record temperatures", "EVTYPE2"]<-"?"
df[df$EVTYPE == "record temperature", "EVTYPE2"]<-"?"
df[df$EVTYPE == "temperature record", "EVTYPE2"]<-"?"
df[df$EVTYPE == "apache county", "EVTYPE2"]<-"?"
df[df$EVTYPE == "glaze", "EVTYPE2"]<-"?"
df[df$EVTYPE == "heavy mix", "EVTYPE2"]<-"?"
df[df$EVTYPE == "monthly temperature", "EVTYPE2"]<-"?"
df[df$EVTYPE == "drowning", "EVTYPE2"]<-"?"
df[df$EVTYPE == "southeast", "EVTYPE2"]<-"?"
df[df$EVTYPE == "hypothermia", "EVTYPE2"]<-"hypothermia/exposure"Coding is tranlated for easier cost computation: “h” or “H” means 10^2, “k” or “K” means 10^3, “m” or “M” means 10^6 and “b” or “B” means 10^9.
df[df$PROPDMGEXP == "h", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "h", "PROPDMG"])) * 100
df[df$PROPDMGEXP == "k", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "k", "PROPDMG"])) * 1000
df[df$PROPDMGEXP == "m", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "m", "PROPDMG"])) * 10^6
df[df$PROPDMGEXP == "b", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "b", "PROPDMG"])) * 10^9
df[df$PROPDMGEXP == "1", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "1", "PROPDMG"])) * 10^1
df[df$PROPDMGEXP == "2", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "2", "PROPDMG"])) * 10^2
df[df$PROPDMGEXP == "3", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "3", "PROPDMG"])) * 10^3
df[df$PROPDMGEXP == "4", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "4", "PROPDMG"])) * 10^4
df[df$PROPDMGEXP == "5", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "5", "PROPDMG"])) * 10^5
df[df$PROPDMGEXP == "6", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "6", "PROPDMG"])) * 10^6
df[df$PROPDMGEXP == "7", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "7", "PROPDMG"])) * 10^7
df[df$PROPDMGEXP == "8", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "8", "PROPDMG"])) * 10^8
df[df$PROPDMGEXP == "9", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "9", "PROPDMG"])) * 10^9
df[df$PROPDMGEXP == "10", "newPROPDMG"] <- as.numeric(as.character(df[df$PROPDMGEXP == "10", "PROPDMG"])) * 10^10
df[df$CROPDMGEXP == "h", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "h", "CROPDMG"])) * 100
df[df$CROPDMGEXP == "k", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "k", "CROPDMG"])) * 1000
df[df$CROPDMGEXP == "m", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "m", "CROPDMG"])) * 10^6
df[df$CROPDMGEXP == "b", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "b", "CROPDMG"])) * 10^9
df[df$CROPDMGEXP == "1", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "1", "CROPDMG"])) * 10^1
df[df$CROPDMGEXP == "2", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "2", "CROPDMG"])) * 10^2
df[df$CROPDMGEXP == "3", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "3", "CROPDMG"])) * 10^3
df[df$CROPDMGEXP == "4", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "4", "CROPDMG"])) * 10^4
df[df$CROPDMGEXP == "5", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "5", "CROPDMG"])) * 10^5
df[df$CROPDMGEXP == "6", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "6", "CROPDMG"])) * 10^6
df[df$CROPDMGEXP == "7", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "7", "CROPDMG"])) * 10^7
df[df$CROPDMGEXP == "8", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "8", "CROPDMG"])) * 10^8
df[df$CROPDMGEXP == "9", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "9", "CROPDMG"])) * 10^9
df[df$CROPDMGEXP == "10", "newCROPDMG"] <- as.numeric(as.character(df[df$CROPDMGEXP == "10", "CROPDMG"])) * 10^10Results
1. Across the United States, which types of events are most harmful with respect to population health?
Total number of deaths across all instances of a type of event, in rank order:
totfatal <- aggregate(as.numeric(as.character(FATALITIES))~EVTYPE2,data=df,FUN=sum)
colnames(totfatal)<- c("EventType", "TotFatalities")
totfatal$rank <- rank(-totfatal$TotFatalities)
totfatal <- totfatal[order(totfatal$rank,decreasing=F), ]
totfatal## EventType TotFatalities rank
## 26 tornado 5658 1.0
## 30 warm weather 2957 2.0
## 9 flood 1556 3.0
## 33 wind 1148 4.0
## 13 hazardous surf 1013 5.0
## 17 lightning 817 6.0
## 6 cold weather 514 7.0
## 2 avalanche 225 8.0
## 24 snow 149 9.0
## 14 hurricane 135 10.0
## 21 rain 115 11.0
## 16 ice 104 12.0
## 3 blizzard 101 13.0
## 32 wildfire 90 14.0
## 10 fog 81 15.0
## 27 tropical storm 66 16.0
## 1 ? 50 17.0
## 19 mud/rock slide 44 18.0
## 4 blowing dust 24 19.0
## 12 hail 20 20.0
## 31 waterspout 6 21.0
## 22 sleet 5 22.0
## 20 other 3 23.0
## 5 coastal storm 1 25.5
## 15 hypothermia/exposure 1 25.5
## 18 mild pattern 1 25.5
## 25 thunderstorm 1 25.5
## 7 dam failure 0 30.5
## 8 excessively dry 0 30.5
## 11 gustnado 0 30.5
## 23 smoke 0 30.5
## 28 volcanic eruption 0 30.5
## 29 wall cloud/funnel cloud 0 30.5maxfatalevent <- totfatal[totfatal$rank == 1, "EventType"]
maxfatalcount <- totfatal[totfatal$rank == 1, "TotFatalities"]
secfatalevent <- totfatal[totfatal$rank == 2, "EventType"]
secfatalcount <- totfatal[totfatal$rank == 2, "TotFatalities"]Total number of injuries across all instances of a type of event, in rank order:
totinjur <- aggregate(as.numeric(as.character(INJURIES))~EVTYPE2,data=df,FUN=sum)
colnames(totinjur)<- c("EventType", "TotInjuries")
totinjur$rank <- rank(-totinjur$TotInjuries)
totinjur <- totinjur[order(totinjur$rank,decreasing=F), ]
totinjur## EventType TotInjuries rank
## 26 tornado 91364 1.0
## 33 wind 11235 2.0
## 30 warm weather 8833 3.0
## 9 flood 8678 4.0
## 17 lightning 5232 5.0
## 16 ice 2186 6.0
## 32 wildfire 1608 7.0
## 12 hail 1467 8.0
## 14 hurricane 1328 9.0
## 13 hazardous surf 1264 10.0
## 24 snow 1161 11.0
## 10 fog 1077 12.0
## 6 cold weather 934 13.0
## 3 blizzard 806 14.0
## 4 blowing dust 483 15.0
## 27 tropical storm 383 16.0
## 21 rain 333 17.0
## 1 ? 280 18.0
## 2 avalanche 170 19.0
## 31 waterspout 72 20.0
## 19 mud/rock slide 60 21.0
## 25 thunderstorm 39 22.0
## 20 other 28 23.0
## 22 sleet 15 24.0
## 7 dam failure 7 25.0
## 8 excessively dry 4 26.0
## 29 wall cloud/funnel cloud 3 27.0
## 5 coastal storm 0 30.5
## 11 gustnado 0 30.5
## 15 hypothermia/exposure 0 30.5
## 18 mild pattern 0 30.5
## 23 smoke 0 30.5
## 28 volcanic eruption 0 30.5maxinjurevent <- totinjur[totinjur$rank == 1, "EventType"]
maxinjurcount <- totinjur[totinjur$rank == 1, "TotInjuries"]
secinjurevent <- totinjur[totinjur$rank == 2, "EventType"]
secinjurcount <- totinjur[totinjur$rank == 2, "TotInjuries"]ggplot(totinjur[1:10,], aes(EventType, TotInjuries, fill=EventType)) + geom_bar(stat = "identity")+ ylab("Total Number of Injuries")+ xlab("Event Type")+ ggtitle("Top Ten Types of Events Causing Injuries")+ theme(axis.text.x = element_text(angle=30, vjust=0.5))
For harm done per event of a given type rather than across all instances of the type of event, the results are different. Some events are rarer but also more harmful to people’s health when they occur.
Mean number of deaths by type of event, in rank order:
meanfatal <- aggregate(as.numeric(as.character(FATALITIES))~EVTYPE2,data=df,FUN=mean)
colnames(meanfatal)<- c("EventType", "MeanFatalities")
meanfatal$rank <- rank(-meanfatal$MeanFatalities)
meanfatal <- meanfatal[order(meanfatal$rank,decreasing=F), ]
meanfatal## EventType MeanFatalities rank
## 30 warm weather 1.0655855856 1.0
## 5 coastal storm 1.0000000000 2.5
## 15 hypothermia/exposure 1.0000000000 2.5
## 2 avalanche 0.5813953488 4.0
## 14 hurricane 0.4655172414 5.0
## 13 hazardous surf 0.3443235894 6.0
## 27 tropical storm 0.0946915352 7.0
## 26 tornado 0.0932324880 8.0
## 19 mud/rock slide 0.0655737705 9.0
## 17 lightning 0.0517809608 10.0
## 22 sleet 0.0515463918 11.0
## 6 cold weather 0.0480418731 12.0
## 10 fog 0.0429936306 13.0
## 4 blowing dust 0.0408163265 14.0
## 3 blizzard 0.0368075802 15.0
## 16 ice 0.0286896552 16.0
## 32 wildfire 0.0212264151 17.0
## 9 flood 0.0180778883 18.0
## 18 mild pattern 0.0147058824 19.0
## 20 other 0.0133928571 20.0
## 25 thunderstorm 0.0109890110 21.0
## 21 rain 0.0093291150 22.0
## 24 snow 0.0084534211 23.0
## 1 ? 0.0039252630 24.0
## 33 wind 0.0032807874 25.0
## 31 waterspout 0.0015523933 26.0
## 12 hail 0.0000689767 27.0
## 7 dam failure 0.0000000000 30.5
## 8 excessively dry 0.0000000000 30.5
## 11 gustnado 0.0000000000 30.5
## 23 smoke 0.0000000000 30.5
## 28 volcanic eruption 0.0000000000 30.5
## 29 wall cloud/funnel cloud 0.0000000000 30.5maxfatalevent <- meanfatal[meanfatal$rank == 1, "EventType"]
maxfatalcount <- meanfatal[meanfatal$rank == 1, "MeanFatalities"]
secfatalevent <- meanfatal[meanfatal$rank == 2, "EventType"]
secfatalcount <- meanfatal[meanfatal$rank == 2, "MeanFatalities"]Mean number of injuries by type of event, in rank order:
meaninjur <- aggregate(as.numeric(as.character(INJURIES))~EVTYPE2,data=df,FUN=mean)
colnames(meaninjur)<- c("EventType", "MeanInjuries")
meaninjur$rank <- rank(-meaninjur$MeanInjuries)
meaninjur <- meaninjur[order(meaninjur$rank,decreasing=F), ]
meaninjur## EventType MeanInjuries rank
## 14 hurricane 4.5793103448 1.0
## 30 warm weather 3.1830630631 2.0
## 26 tornado 1.5054954109 3.0
## 4 blowing dust 0.8214285714 4.0
## 16 ice 0.6030344828 5.0
## 10 fog 0.5716560510 6.0
## 27 tropical storm 0.5494978479 7.0
## 2 avalanche 0.4392764858 8.0
## 13 hazardous surf 0.4296397009 9.0
## 25 thunderstorm 0.4285714286 10.0
## 32 wildfire 0.3792452830 11.0
## 17 lightning 0.3316009634 12.0
## 3 blizzard 0.2937317784 13.0
## 22 sleet 0.1546391753 14.0
## 7 dam failure 0.1272727273 15.0
## 20 other 0.1250000000 16.0
## 9 flood 0.1008225672 17.0
## 19 mud/rock slide 0.0894187779 18.0
## 6 cold weather 0.0872978783 19.0
## 24 snow 0.0658686032 20.0
## 33 wind 0.0321077058 21.0
## 21 rain 0.0270138720 22.0
## 1 ? 0.0219814728 23.0
## 31 waterspout 0.0186287193 24.0
## 12 hail 0.0050594407 25.0
## 8 excessively dry 0.0015847861 26.0
## 29 wall cloud/funnel cloud 0.0004285102 27.0
## 5 coastal storm 0.0000000000 30.5
## 11 gustnado 0.0000000000 30.5
## 15 hypothermia/exposure 0.0000000000 30.5
## 18 mild pattern 0.0000000000 30.5
## 23 smoke 0.0000000000 30.5
## 28 volcanic eruption 0.0000000000 30.5maxinjurevent <- meaninjur[meaninjur$rank == 1, "EventType"]
maxinjurcount <- meaninjur[meaninjur$rank == 1, "MeanInjuries"]
secinjurevent <- meaninjur[meaninjur$rank == 2, "EventType"]
secinjurcount <- meaninjur[meaninjur$rank == 2, "MeanInjuries"]2. Across the United States, which types of events have the greatest economic consequences?
Total damage to property:
totpropdamage <- aggregate(newPROPDMG~EVTYPE2,data=df,FUN=sum)
colnames(totpropdamage)<- c("EventType", "TotPropDamage")
totpropdamage$rank <- rank(-totpropdamage$TotPropDamage)
totpropdamage <- totpropdamage[order(totpropdamage$rank,decreasing=F), ]
totpropdamage## EventType TotPropDamage rank
## 8 flood 168219697920 1
## 13 hurricane 84756180010 2
## 24 tornado 58552152230 3
## 12 hazardous surf 48103512550 4
## 31 wind 16056882630 5
## 11 hail 16022837220 6
## 30 wildfire 8501628500 7
## 25 tropical storm 7714390550 8
## 14 ice 3970045510 9
## 19 rain 3237520690 10
## 23 thunderstorm 1208461550 11
## 7 excessively dry 1046106000 12
## 22 snow 1011999740 13
## 15 lightning 935737780 14
## 3 blizzard 664913950 15
## 17 mud/rock slide 379548100 16
## 5 cold weather 273889900 17
## 29 waterspout 60737700 18
## 9 fog 25011500 19
## 6 dam failure 10351000 20
## 1 ? 9936740 21
## 28 warm weather 9665700 22
## 18 other 7349600 23
## 4 blowing dust 6337630 24
## 2 avalanche 3721800 25
## 26 volcanic eruption 500000 26
## 27 wall cloud/funnel cloud 199600 27
## 20 sleet 145000 28
## 10 gustnado 102050 29
## 21 smoke 100000 30
## 16 mild pattern 61100 31maxevent <- totpropdamage[totpropdamage$rank == 1, "EventType"]
maxcount <- totpropdamage[totpropdamage$rank == 1, "TotPropDamage"]
secevent <- totpropdamage[totpropdamage$rank == 2, "EventType"]
seccount <- totpropdamage[totpropdamage$rank == 2, "TotPropDamage"]Total damage to crops:
totcropdamage <- aggregate(newCROPDMG~EVTYPE2,data=df,FUN=sum)
colnames(totcropdamage)<- c("EventType", "TotCropDamage")
totcropdamage$rank <- rank(-totcropdamage$TotCropDamage)
totcropdamage <- totcropdamage[order(totcropdamage$rank,decreasing=F), ]
totcropdamage## EventType TotCropDamage rank
## 7 excessively dry 13972566000 1
## 8 flood 12388597200 2
## 14 ice 6722945300 3
## 13 hurricane 5515292800 4
## 11 hail 3111712850 5
## 30 wind 1960859050 6
## 5 cold weather 1335623000 7
## 18 rain 1007662800 8
## 27 warm weather 898869280 9
## 24 tropical storm 694896000 10
## 23 tornado 417461360 11
## 29 wildfire 403281630 12
## 6 dam failure 296285000 13
## 21 snow 134663100 14
## 3 blizzard 112060000 15
## 12 hazardous surf 48107500 16
## 16 mud/rock slide 20017000 17
## 22 thunderstorm 18205000 18
## 15 lightning 12097090 19
## 4 blowing dust 3600000 20
## 1 ? 50000 21
## 17 other 15000 22
## 10 gustnado 1550 23
## 2 avalanche 0 27
## 9 fog 0 27
## 19 sleet 0 27
## 20 smoke 0 27
## 25 volcanic eruption 0 27
## 26 wall cloud/funnel cloud 0 27
## 28 waterspout 0 27maxevent <- totcropdamage[totcropdamage$rank == 1, "EventType"]
maxcount <- totcropdamage[totcropdamage$rank == 1, "TotCropDamage"]
secevent <- totcropdamage[totcropdamage$rank == 2, "EventType"]
seccount <- totcropdamage[totcropdamage$rank == 2, "TotCropDamage"]