Reproducible Research: Peer Assessment 2
Explore Impact of Severe Weather on Public Health & Economic Conditions
Synopsis
- In this report, we try to gauge the impact of Natural Calamities on Human Life and analyse the Ecomonic Damage caused by them
- We tap into U.S. National Oceanic and Atmospheric Administration’s (NOAA) storm database, which captures Major Storms & Other Severe Weather Condition Events in the US, from 1950 to 2011
- Information retrieved is : Various Event Types, Date Of Occurrance, States affected, Counts of Fatalities & Injuries and Property & Crop Damage in Dollars
- To get the Accurate information, Event Types are Streamlined, wherever applicable, as there are 985 Event Types reported versus only 48 Official Event Types published by NOAA
- The Property & Crop Damage Units are converted into proper numerical values
- As part of the Analysis, we tried to find out Top Fatalites/Injuries By Event, Top 5 Worst Hit States, Top Property/Crop Damage By Event and Most Distatrous Years hit by Worst Calamities
- From the Analysis, we conclude that TORNADOs cause maximum Fatalities/Injuries, FLOOD is the singlemost cause of severe Property Damage (122 Billion Dollars) and 2006 was the Worst Hit Year with 115 Billion Dollars lost
- For Complete Analysis, please see the CONCLUSIONS section, at the bottom of this report
Data Processing
The data for this assignment is retrieved from the U.S. National Oceanic and Atmospheric Administration’s (NOAA) storm database.
Here’s the [Storm Data] (https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2) [47Mb]
Here’s some [Storm Data Documentation] (https://d396qusza40orc.cloudfront.net/repdata%2Fpeer2_doc%2Fpd01016005curr.pdf) from National Weather Service.
Loading and PreProcessing the Data
Download the bz2 file from the above URL and load it,
if (!file.exists("repdata-data-StormData.csv.bz2")) {
download.file(url="http://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2",
destfile="repdata-data-StormData.csv.bz2")
}stormData <- read.csv("repdata-data-StormData.csv.bz2",
header=T,
sep=",",
na.strings="",
stringsAsFactors=F)This is how the data appears,
dim(stormData)## [1] 902297 37str(stormData)## 'data.frame': 902297 obs. of 37 variables:
## $ STATE__ : num 1 1 1 1 1 1 1 1 1 1 ...
## $ BGN_DATE : chr "4/18/1950 0:00:00" "4/18/1950 0:00:00" "2/20/1951 0:00:00" "6/8/1951 0:00:00" ...
## $ BGN_TIME : chr "0130" "0145" "1600" "0900" ...
## $ TIME_ZONE : chr "CST" "CST" "CST" "CST" ...
## $ COUNTY : num 97 3 57 89 43 77 9 123 125 57 ...
## $ COUNTYNAME: chr "MOBILE" "BALDWIN" "FAYETTE" "MADISON" ...
## $ STATE : chr "AL" "AL" "AL" "AL" ...
## $ EVTYPE : chr "TORNADO" "TORNADO" "TORNADO" "TORNADO" ...
## $ BGN_RANGE : num 0 0 0 0 0 0 0 0 0 0 ...
## $ BGN_AZI : chr NA NA NA NA ...
## $ BGN_LOCATI: chr NA NA NA NA ...
## $ END_DATE : chr NA NA NA NA ...
## $ END_TIME : chr NA NA NA NA ...
## $ COUNTY_END: num 0 0 0 0 0 0 0 0 0 0 ...
## $ COUNTYENDN: logi NA NA NA NA NA NA ...
## $ END_RANGE : num 0 0 0 0 0 0 0 0 0 0 ...
## $ END_AZI : chr NA NA NA NA ...
## $ END_LOCATI: chr NA NA NA NA ...
## $ LENGTH : num 14 2 0.1 0 0 1.5 1.5 0 3.3 2.3 ...
## $ WIDTH : num 100 150 123 100 150 177 33 33 100 100 ...
## $ F : int 3 2 2 2 2 2 2 1 3 3 ...
## $ MAG : num 0 0 0 0 0 0 0 0 0 0 ...
## $ FATALITIES: num 0 0 0 0 0 0 0 0 1 0 ...
## $ INJURIES : num 15 0 2 2 2 6 1 0 14 0 ...
## $ PROPDMG : num 25 2.5 25 2.5 2.5 2.5 2.5 2.5 25 25 ...
## $ PROPDMGEXP: chr "K" "K" "K" "K" ...
## $ CROPDMG : num 0 0 0 0 0 0 0 0 0 0 ...
## $ CROPDMGEXP: chr NA NA NA NA ...
## $ WFO : chr NA NA NA NA ...
## $ STATEOFFIC: chr NA NA NA NA ...
## $ ZONENAMES : chr NA NA NA NA ...
## $ LATITUDE : num 3040 3042 3340 3458 3412 ...
## $ LONGITUDE : num 8812 8755 8742 8626 8642 ...
## $ LATITUDE_E: num 3051 0 0 0 0 ...
## $ LONGITUDE_: num 8806 0 0 0 0 ...
## $ REMARKS : chr NA NA NA NA ...
## $ REFNUM : num 1 2 3 4 5 6 7 8 9 10 ...Question 1: Across the United States, which types of events (as indicated in the EVTYPE variable) are most harmful with respect to population health?
We shall solve this question in 3 ways,
A) What are the “Top Fatalities By Event” And “Top Injuries By Event”,
Include dplyr package. Summarize {fatalities} and {injuries} by {EVENT TYPES},
library(dplyr)##
## Attaching package: 'dplyr'
##
## The following object is masked from 'package:stats':
##
## filter
##
## The following objects are masked from 'package:base':
##
## intersect, setdiff, setequal, unionstormCasualties <- stormData %>%
group_by(EVTYPE) %>% summarize(fatalities = sum(FATALITIES, na.rm=TRUE),
injuries = sum(INJURIES, na.rm=TRUE))Consider only those Casualties where there are both {fatalities} and {injuries} present. The proportion of, either, {fatalities} with zero {injuries} or {injuries} with 0 fatalities, is considered to be small and not taken into consideration for this analysis.
tidyStormCasualties <- subset(stormCasualties, fatalities > 0 & injuries > 0)View Casualties Data,
dim(tidyStormCasualties)## [1] 106 3Notice that there are lot of anomalies in the {EVENT TYPE} variable. The NOAA Database officially defines only 48 EventTypes. Let’s fix the {EVENT TYPE} variable, by writing a function fixEvtype. We will be calling this function multiple times.
fixEvtype <- function(x) {
x$EVTYPE <- sub("^COLD$", "COLD/WIND CHILL", x$EVTYPE)
x$EVTYPE <- sub("^EXTREME COLD$", "EXTREME COLD/WIND CHILL", x$EVTYPE)
x$EVTYPE <- sub("^EXTREME HEAT$", "EXCESSIVE HEAT", x$EVTYPE)
x$EVTYPE <- sub("^EXTREME WINDCHILL$", "EXTREME COLD/WIND CHILL", x$EVTYPE)
x$EVTYPE <- sub("^FLASH FLOODING$", "FLASH FLOOD", x$EVTYPE)
x$EVTYPE <- sub("^FLOOD/FLASH FLOOD$", "FLOOD", x$EVTYPE)
x$EVTYPE <- sub("^FLOODING$", "FLOOD", x$EVTYPE)
x$EVTYPE <- sub("^FREEZING DRIZZLE$", "FROST/FREEZE", x$EVTYPE)
x$EVTYPE <- sub("^FREEZING RAIN$", "FROST/FREEZE", x$EVTYPE)
x$EVTYPE <- sub("^FROST$", "FROST/FREEZE", x$EVTYPE)
x$EVTYPE <- sub("^GUSTY WIND$", "HIGH WIND", x$EVTYPE)
x$EVTYPE <- sub("^GUSTY WINDS$", "HIGH WIND", x$EVTYPE)
x$EVTYPE <- sub("^HEAT WAVE$", "HEAT", x$EVTYPE)
x$EVTYPE <- sub("^HEAVY SURF$", "HIGH SURF", x$EVTYPE)
x$EVTYPE <- sub("^HEAVY SURF/HIGH SURF$", "HIGH SURF", x$EVTYPE)
x$EVTYPE <- sub("^HIGH WIND AND SEAS$", "MARINE HIGH WIND", x$EVTYPE)
x$EVTYPE <- sub("^HIGH WINDS$", "HIGH WIND", x$EVTYPE)
x$EVTYPE <- sub("^HURRICANE$", "HURRICANE/TYPHOON", x$EVTYPE)
x$EVTYPE <- sub("^HURRICANE ERIN$", "HURRICANE/TYPHOON", x$EVTYPE)
x$EVTYPE <- sub("^HURRICANE OPAL$", "HURRICANE/TYPHOON", x$EVTYPE)
x$EVTYPE <- sub("^High Surf$", "HIGH SURF", x$EVTYPE)
x$EVTYPE <- sub("^ICE$", "ICE STORM", x$EVTYPE)
x$EVTYPE <- sub("^LANDSLIDES$", "LANDSLIDE", x$EVTYPE)
x$EVTYPE <- sub("^MARINE TSTM WIND$", "MARINE THUNDERSTORM WIND", x$EVTYPE)
x$EVTYPE <- sub("^Marine Accident$", "MARINE ACCIDENT", x$EVTYPE)
x$EVTYPE <- sub("^Mudslide$", "MUDSLIDE", x$EVTYPE)
x$EVTYPE <- sub("^RECORD HEAT$", "EXCESSIVE HEAT", x$EVTYPE)
x$EVTYPE <- sub("^RIP CURRENTS$", "RIP CURRENT", x$EVTYPE)
x$EVTYPE <- sub("^ROUGH SURF$", "HIGH SURF", x$EVTYPE)
x$EVTYPE <- sub("^STORM SURGE$", "STORM SURGE/TIDE", x$EVTYPE)
x$EVTYPE <- sub("^STRONG WINDS$", "STRONG WIND", x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORM$", "THUNDERSTORM WIND", x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORM WINDS$", "THUNDERSTORM WIND", x$EVTYPE)
x$EVTYPE <- sub("^TROPICAL STORM GORDON$", "TROPICAL STORM", x$EVTYPE)
x$EVTYPE <- sub("^TSTM WIND$", "THUNDERSTORM WIND", x$EVTYPE)
x$EVTYPE <- sub("^TSTM WIND/HAIL$", "THUNDERSTORM WIND", x$EVTYPE)
x$EVTYPE <- sub("^UNSEASONABLY WARM$", "EXCESSIVE HEAT", x$EVTYPE)
x$EVTYPE <- sub("^WATERSPOUT/TORNADO$", "WATERSPOUT", x$EVTYPE)
x$EVTYPE <- sub("^WILD FIRES$", "WILDFIRE", x$EVTYPE)
x$EVTYPE <- sub("^WILD/FOREST FIRE$", "WILDFIRE", x$EVTYPE)
x$EVTYPE <- sub("^WIND$", "STRONG WIND", x$EVTYPE)
x$EVTYPE <- sub("^WINDS$", "STRONG WIND", x$EVTYPE)
x$EVTYPE <- sub("^WINTER STORM HIGH WINDS$", "WINTER STORM", x$EVTYPE)
x$EVTYPE <- sub("^WINTER STORMS$", "WINTER STORM", x$EVTYPE)
x$EVTYPE <- sub("^WINTER WEATHER/MIX$", "WINTER WEATHER", x$EVTYPE)
x$EVTYPE <- sub("^WINTRY MIX$", "WINTER WEATHER", x$EVTYPE)
x$EVTYPE <- sub("^blowing snow$", "BLOWING SNOW", x$EVTYPE)
x$EVTYPE <- sub("^COASTAL FLOODING$", "COASTAL FLOOD",x$EVTYPE)
x$EVTYPE <- sub("^DUST STORM/HIGH WINDS$", "DUST STORM", x$EVTYPE)
x$EVTYPE <- sub("^FLASH FLOOD/FLOOD$", "FLASH FLOOD",x$EVTYPE)
x$EVTYPE <- sub("^FLASH FLOODING/FLOOD$", "FLASH FLOOD",x$EVTYPE)
x$EVTYPE <- sub("^Frost/Freeze$", "FROST/FREEZE",x$EVTYPE)
x$EVTYPE <- sub("^HAIL 100$", "HAIL",x$EVTYPE)
x$EVTYPE <- sub("^HAIL/WIND$", "HAIL",x$EVTYPE)
x$EVTYPE <- sub("^HAIL/WINDS$", "HAIL",x$EVTYPE)
x$EVTYPE <- sub("^HEAVY RAINS$", "HEAVY RAIN",x$EVTYPE)
x$EVTYPE <- sub("^HEAVY RAINS/FLOODING$", "HEAVY RAIN",x$EVTYPE)
x$EVTYPE <- sub("^HEAVY SNOW/HIGH WINDS & FLOOD$", "HEAVY SNOW",x$EVTYPE)
x$EVTYPE <- sub("^HIGH WINDS HEAVY RAINS$", "HIGH WIND",x$EVTYPE)
x$EVTYPE <- sub("^HIGH WINDS/COLD$", "HIGH WIND",x$EVTYPE)
x$EVTYPE <- sub("^HURRICANE FELIX$", "HURRICANE/TYPHOON",x$EVTYPE)
x$EVTYPE <- sub("^HURRICANE OPAL/HIGH WINDS$", "HURRICANE/TYPHOON", x$EVTYPE)
x$EVTYPE <- sub("^Heavy Rain/High Surf$", "HEAVY RAIN",x$EVTYPE)
x$EVTYPE <- sub("^RIVER FLOODING$", "RIVER FLOOD", x$EVTYPE)
x$EVTYPE <- sub("^River Flooding$", "RIVER FLOOD",x$EVTYPE)
x$EVTYPE <- sub("^SEVERE THUNDERSTORM WINDS$", "THUNDERSTORM WIND", x$EVTYPE)
x$EVTYPE <- sub("^SEVERE THUNDERSTORMS$", "THUNDERSTORM WIND", x$EVTYPE)
x$EVTYPE <- sub("^SMALL HAIL$", "HAIL",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORM WINDS HAIL$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORM WINDS LIGHTNING$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORM WINDS/ FLOOD$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORM WINDS/HAIL$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORM WINDSS$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORMS$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORMS WIND$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^THUNDERSTORMS WINDS$", "THUNDERSTORM WIND",x$EVTYPE)
x$EVTYPE <- sub("^TORNADO F0$", "TORNADO",x$EVTYPE)
x$EVTYPE <- sub("^TORNADOES, TSTM WIND, HAIL$", "TORNADO",x$EVTYPE)
x$EVTYPE <- sub("^TROPICAL STORM DEAN$", "TROPICAL STORM",x$EVTYPE)
x$EVTYPE <- sub("^TROPICAL STORM JERRY$", "TROPICAL STORM",x$EVTYPE)
x$EVTYPE <- sub("^URBAN FLOOD$", "FLOOD",x$EVTYPE)
x$EVTYPE <- sub("^URBAN FLOODING$", "FLOOD",x$EVTYPE)
x$EVTYPE <- sub("^URBAN/SML$", "STREAM FLD", x$EVTYPE)
x$EVTYPE <- sub("^WILD/FOREST FIRES$", "WILDFIRE",x$EVTYPE)
x$EVTYPE <- sub("^WILDFIRES$", "WILDFIRE",x$EVTYPE)
return(x)
}Now let’s apply this function,
tidyStormCasualties2 <- tidyStormCasualties
tidyStormCasualties2 <- fixEvtype(tidyStormCasualties2)
tidyStormCasualties3 <- tidyStormCasualties2 %>%
group_by(EVTYPE) %>% summarize(fatalities = sum(fatalities, na.rm=TRUE),
injuries = sum(injuries, na.rm=TRUE))View Casualties Data again,
dim(tidyStormCasualties3)## [1] 62 3summary(tidyStormCasualties3$fatalities)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.00 3.25 16.50 241.00 129.20 5633.00Now, let’s find out those {EVENTS} where {fatalities} are Greater Than Mean,
topTidyStormCasualties <- subset(tidyStormCasualties3,
fatalities > mean(fatalities))
topTidyStormCasualties$EVTYPE <- factor(topTidyStormCasualties$EVTYPE)Let’s shift our focus to {injuries} now. The fix for {EVENT TYPE} has already been applied.
View Injuries Data,
dim(tidyStormCasualties3)## [1] 62 3summary(tidyStormCasualties3$injuries)## Min. 1st Qu. Median Mean 3rd Qu. Max.
## 1.00 14.25 56.50 2261.00 687.20 91350.00Get those {EVENTS} where {injuries} are Greater Than Mean,
topTidyStormInjuries <- subset(tidyStormCasualties3,
injuries > mean(injuries))
topTidyStormInjuries$EVTYPE <- factor(topTidyStormInjuries$EVTYPE)B) Which “Top 5 States” are “Worst Hit” by most detrimental EVENTS,
Let’s find out the Top 5 Worst Hit States first.
stateCasualties <- stormData %>%
group_by(STATE) %>% summarize(fatalities = sum(FATALITIES, na.rm=TRUE),
injuries = sum(INJURIES, na.rm=TRUE),
casualties = fatalities + injuries)
names(stateCasualties) <- c("state","fatalities","injuries","casualties")
top5stateCasualties <- head(arrange(stateCasualties, desc(casualties)),5)View Top 5 Worst Hit States,
top5stateCasualties[,1]## Source: local data frame [5 x 1]
##
## state
## 1 TX
## 2 MO
## 3 AL
## 4 OH
## 5 MSGet the Storm Data only for these 5 States now. Summarize it by {STATE} and {EVTYPE},
top5stateStormData <- subset(stormData, STATE %in% top5stateCasualties$state)
top5stateStormCasualties <- top5stateStormData %>%
group_by(STATE, EVTYPE) %>% summarize(fatalities = sum(FATALITIES, na.rm=TRUE),
injuries = sum(INJURIES, na.rm=TRUE),
casualties = fatalities + injuries)Consider only those Casualties where there are both {fatalities} and {injuries},
top5stateTidyStormCasualties <- subset(top5stateStormCasualties, fatalities > 0 & injuries > 0)View State Casualties Data,
dim(top5stateTidyStormCasualties)## [1] 80 5Apply the {EVENT TYPE} fix now,
top5stateTidyStormCasualties2 <- top5stateTidyStormCasualties
top5stateTidyStormCasualties2 <- fixEvtype(top5stateTidyStormCasualties2)
top5stateTidyStormCasualties3 <- top5stateTidyStormCasualties2 %>%
group_by(STATE, EVTYPE) %>% summarize(fatalities = sum(fatalities, na.rm=TRUE),
injuries = sum(injuries, na.rm=TRUE),
casualties = fatalities + injuries)View State Casualties Data again,
dim(top5stateTidyStormCasualties3)## [1] 65 5Let’s select only Top 5 EVENTS per State,
top5state5stormCasualtyData <- filter(top5stateTidyStormCasualties3, STATE == "ZZ")
for (i in top5stateCasualties$state) {
a <- head(arrange(filter(top5stateTidyStormCasualties3, STATE == i), desc(casualties)), 5)
top5state5stormCasualtyData <- rbind(top5state5stormCasualtyData, a)
}
top5state5stormCasualtyData$EVTYPE <- factor(top5state5stormCasualtyData$EVTYPE)C) Which are the Most Disastrous Years hit by Worst Casualties,
We shall use BGN_DATE for our Date-Calculations. Let’s extract the Year from the BGN_DATE and build summary,
CasualtyDamageData <- subset(stormData, FATALITIES > 0 & INJURIES > 0)
CasualtyDamageData <- CasualtyDamageData %>%
mutate(beginYear = format(as.Date(BGN_DATE,"%m/%d/%Y"), "%Y"))
yearlyCasualtyDmgSummary <- CasualtyDamageData %>%
group_by(beginYear) %>% summarize(fatalities = sum(FATALITIES, na.rm=TRUE),
injuries = sum(INJURIES, na.rm=TRUE),
casualties = fatalities + injuries)Now we turn our attention to second question,
Question 2: Across the United States, which types of events have the greatest economic consequences?
We shall solve this question in 2 ways,
A) What is the “Top Property Damage By Event” And “Top Crop Damage By Event”,
The economic data is stored in 4 variables, ie. {PROPDMG}, {PROPDMGEXP}, {CROPDMG}, {CROPDMGEXP}.
The numerical values are stored in {PROPDMG} and {CROPDMG}. The units are stored in {PROPDMGEXP} and {CROPDMGEXP}.
Let’s see how the units are distributed,
table(stormData$PROPDMGEXP)##
## - ? + 0 1 2 3 4 5 6
## 1 8 5 216 25 13 4 4 28 4
## 7 8 B h H K m M
## 5 1 40 1 6 424665 7 11330table(stormData$CROPDMGEXP)##
## ? 0 2 B k K m M
## 7 19 1 9 21 281832 1 1994For the sake of this analysis, we shall take only those observations where {PROPDMG} is present and {CROPDMG} is present.
stormDamage <- subset(stormData, PROPDMG > 0 & !is.na(PROPDMGEXP) &
CROPDMG > 0 & !is.na(CROPDMGEXP))Here are the units one more time,
table(stormDamage$PROPDMGEXP)##
## 0 3 5 B K m M
## 4 1 2 15 14848 1 1363table(stormDamage$CROPDMGEXP)##
## 0 B k K m M
## 11 3 21 14906 1 1292Let’s define 2 new variables, ie. {propDmgMultiplier}, {cropDmgMultiplier}, to calculate numerical values of units,
The only valid units we are considering, are,
- (k,K) = Thousands
- (m,M) = Millions
- (b,B) = Billions
- = Hundreds
stormDamage <- stormDamage %>%
mutate(propDmgMultiplier = 0, cropDmgMultiplier = 0)
for (i in 1:nrow(stormDamage)) {
if (stormDamage[i,]$PROPDMGEXP %in% c("k","K")) {
stormDamage[i,]$propDmgMultiplier = 1000
} else if (stormDamage[i,]$PROPDMGEXP %in% c("m","M")) {
stormDamage[i,]$propDmgMultiplier = 1000000
} else if (stormDamage[i,]$PROPDMGEXP %in% c("b","B")) {
stormDamage[i,]$propDmgMultiplier = 1000000000
} else if (stormDamage[i,]$PROPDMGEXP == "H") {
stormDamage[i,]$propDmgMultiplier = 100
} else {
stormDamage[i,]$propDmgMultiplier = 0
}
}
for (i in 1:nrow(stormDamage)) {
if (stormDamage[i,]$CROPDMGEXP %in% c("k","K")) {
stormDamage[i,]$cropDmgMultiplier = 1000
} else if (stormDamage[i,]$CROPDMGEXP %in% c("m","M")) {
stormDamage[i,]$cropDmgMultiplier = 1000000
} else if (stormDamage[i,]$CROPDMGEXP %in% c("b","B")) {
stormDamage[i,]$cropDmgMultiplier = 1000000000
} else if (stormDamage[i,]$CROPDMGEXP == "H") {
stormDamage[i,]$cropDmgMultiplier = 100
} else {
stormDamage[i,]$cropDmgMultiplier = 0
}
}Now, it’s time to summarize Property and Crop damage by EVTYPE,
propDmgSummary <- stormDamage %>%
group_by(EVTYPE) %>%
summarize(propertyDamage = sum(PROPDMG * propDmgMultiplier),
cropDamage = sum(CROPDMG * cropDmgMultiplier))
dim(propDmgSummary)## [1] 96 3Let’s fix EVENT TYPES now,
propDmgSummary2 <- propDmgSummary
propDmgSummary2 <- fixEvtype(propDmgSummary2)
propDmgSummary3 <- propDmgSummary2 %>%
group_by(EVTYPE) %>% summarize(propertyDamage = sum(propertyDamage),
cropDamage = sum(cropDamage))
dim(propDmgSummary3)## [1] 42 3Get the Top 10 “PropertyDamage” Observations,
topPropDmgSummary <- head(arrange(propDmgSummary3, desc(propertyDamage)),10)
topPropDmgSummary$EVTYPE <- factor(topPropDmgSummary$EVTYPE)Get all “CropDamage” Observations, which are Greater Than MEAN,
topCropDmgSummary <- subset(propDmgSummary3,
cropDamage > mean(cropDamage))
topCropDmgSummary$EVTYPE <- factor(topCropDmgSummary$EVTYPE)B) Which are the Most Disastrous Years for Property & Crop Damage,
We shall use BGN_DATE for our Date-Calculations. Let’s extract the Year from the BGN_DATE and build summary,
stormDamage2 <- stormDamage
stormDamage2 <- stormDamage2 %>%
mutate(beginYear = format(as.Date(BGN_DATE,"%m/%d/%Y"), "%Y"))
yearlyPropDmgSummary <- stormDamage2 %>%
group_by(beginYear) %>%
summarize(propertyDamage = sum(PROPDMG * propDmgMultiplier),
cropDamage = sum(CROPDMG * cropDmgMultiplier))Results
Question 1: Across the United States, which types of events (as indicated in the EVTYPE variable) are most harmful with respect to population health?
We shall answer this question in 3 ways,
A) What are the “Top Fatalities By Event” And “Top Injuries By Event”,
Let’s review the summarized {fatalities} data,
select(arrange(topTidyStormCasualties, desc(fatalities)), EVTYPE, fatalities)## Source: local data frame [10 x 2]
##
## EVTYPE fatalities
## 1 TORNADO 5633
## 2 EXCESSIVE HEAT 2012
## 3 HEAT 1109
## 4 FLASH FLOOD 997
## 5 LIGHTNING 816
## 6 THUNDERSTORM WIND 707
## 7 RIP CURRENT 572
## 8 FLOOD 493
## 9 EXTREME COLD/WIND CHILL 302
## 10 HIGH WIND 288Let’s review the summarized {injuries} data,
select(arrange(topTidyStormInjuries, desc(injuries)), EVTYPE, injuries)## Source: local data frame [6 x 2]
##
## EVTYPE injuries
## 1 TORNADO 91346
## 2 THUNDERSTORM WIND 9460
## 3 FLOOD 6806
## 4 EXCESSIVE HEAT 6747
## 5 LIGHTNING 5230
## 6 HEAT 2409Plot the chart of Top Fatalities By Event Versus Top Injuries By Event,
par(mfrow=c(1,2), oma=c(0,0,2,0))
par(mar=c(13.1, 4.1, 4.1, 2.1), mgp=c(3, 0.5, 0), cex=0.75)
with(topTidyStormCasualties, {
plot(EVTYPE, fatalities, xlab="",
ylab="Number of Fatalities",
type="n",
main="Top Fatalities By Event",
font.main=2, cex.main=1.5, las=2)
points(EVTYPE, fatalities, pch=19, col="black")
lines(EVTYPE, fatalities, lwd=2, col="green")
mtext("Figure 1", outer = TRUE)
})
with(topTidyStormInjuries, {
plot(EVTYPE, injuries, xlab="",
ylab="Number of Injuries",
type="n",
main="Top Injuries By Event",
font.main=2, cex.main=1.5, las=2)
points(EVTYPE, injuries, pch=19, col="black")
lines(EVTYPE, injuries, lwd=2, col="blue")
})
B) Which “Top 5 States” are “Worst Hit” by most detrimental EVENTS,
Let’s review the Worst Hit Top 5 STATES, along with their Top 5 EVENTS.
{casualties} are calculated as {fatalities} + {injuries},
options(dplyr.print_max = 25)
top5state5stormCasualtyData## Source: local data frame [25 x 5]
## Groups: STATE
##
## STATE EVTYPE fatalities injuries casualties
## 1 TX TORNADO 538 8207 8745
## 2 TX FLOOD 49 6338 6387
## 3 TX HEAT 29 774 803
## 4 TX FLASH FLOOD 177 587 764
## 5 TX THUNDERSTORM WIND 48 706 754
## 6 MO TORNADO 388 4330 4718
## 7 MO EXCESSIVE HEAT 190 3525 3715
## 8 MO HEAT 39 660 699
## 9 MO THUNDERSTORM WIND 14 320 334
## 10 MO FLASH FLOOD 72 36 108
## 11 AL TORNADO 617 7929 8546
## 12 AL THUNDERSTORM WIND 40 431 471
## 13 AL LIGHTNING 28 154 182
## 14 AL DENSE FOG 1 71 72
## 15 AL HEAT 9 50 59
## 16 OH TORNADO 191 4438 4629
## 17 OH ICE STORM 2 1652 1654
## 18 OH THUNDERSTORM WIND 46 310 356
## 19 OH LIGHTNING 32 163 195
## 20 OH HEAT 18 75 93
## 21 MS TORNADO 450 6244 6694
## 22 MS THUNDERSTORM WIND 21 246 267
## 23 MS HURRICANE/TYPHOON 16 104 120
## 24 MS LIGHTNING 16 39 55
## 25 MS EXCESSIVE HEAT 24 5 29And here’s their chart,
library(ggplot2)
g <- ggplot(top5state5stormCasualtyData,aes(EVTYPE,casualties,fill=STATE)) +
guides(fill=FALSE) +
geom_bar(stat="identity") +
facet_grid(STATE~.) +
labs(x="", y="Number of Casualties") +
labs(title="Figure 2 : Top 5 Casualty Events By State") +
theme(axis.text.x=element_text(angle=90),
plot.title = element_text(face="bold", size=20),
strip.text.y = element_text(face="bold", size = 12),
strip.background = element_rect(fill="orange"))
print(g)
C) Which are the Most Disastrous Years hit by Worst Casualties,
Extract the Top 5 Years for Worst Casualties,
head(select(arrange(yearlyCasualtyDmgSummary, desc(casualties)), beginYear, casualties),5)## Source: local data frame [5 x 2]
##
## beginYear casualties
## 1 2011 6573
## 2 1974 5622
## 3 1953 5196
## 4 1998 4871
## 5 1965 4413Question 2: Across the United States, which types of events have the greatest economic consequences?
We shall answer this question in 2 ways,
A) What is the “Top Property Damage By Event” And “Top Crop Damage By Event”,
Let’s review the summarized {Property Damage} data,
select(arrange(topPropDmgSummary, desc(propertyDamage)), EVTYPE, propertyDamage)## Source: local data frame [10 x 2]
##
## EVTYPE propertyDamage
## 1 FLOOD 122101960050
## 2 HURRICANE/TYPHOON 37074073000
## 3 RIVER FLOOD 5185675000
## 4 TORNADO 3587241200
## 5 FLASH FLOOD 3260763010
## 6 HAIL 1960933940
## 7 THUNDERSTORM WIND 1771077900
## 8 HIGH WIND 1512136200
## 9 WILDFIRE 1474543000
## 10 TROPICAL STORM 868880400Let’s review the summarized {Crop Damage} data,
select(arrange(topCropDmgSummary, desc(cropDamage)), EVTYPE, cropDamage)## Source: local data frame [8 x 2]
##
## EVTYPE cropDamage
## 1 HURRICANE/TYPHOON 5332242800
## 2 RIVER FLOOD 5056759000
## 3 ICE STORM 5022113500
## 4 FLOOD 4178622950
## 5 HAIL 1879124150
## 6 FLASH FLOOD 1393066400
## 7 DROUGHT 1231755000
## 8 THUNDERSTORM WIND 1087681350Plot the chart of Top Property Damage By Event Versus Top Crop Damage By Event,
par(mfrow=c(1,2), oma=c(0,0,2,0))
par(mar=c(13.1, 4.1, 4.1, 2.1), mgp=c(3, 0.5, 0), cex=0.75)
with(topPropDmgSummary, {
plot(EVTYPE, propertyDamage/1000000000, xlab="",
ylab="Propety Damage in Billion Dollars",
type="n",
main="Top Property Damage By Event",
font.main=2, cex.main=1.5, las=2)
points(EVTYPE, propertyDamage/1000000000, pch=19, col="red")
mtext("Figure 3", outer = TRUE)
})
with(topCropDmgSummary, {
plot(EVTYPE, cropDamage/1000000000, xlab="",
ylab="Crop Damage in Billion Dollars",
type="n",
main="Top Crop Damage By Event",
font.main=2, cex.main=1.5, las=2)
points(EVTYPE, cropDamage/1000000000, pch=19, col="orange")
})
B) Which are the Most Disastrous Years for Property & Crop Damage,
Extract the Top 5 Years for Property & Crop Damage,
head(select(arrange(yearlyPropDmgSummary, desc(propertyDamage)), beginYear, propertyDamage),5)## Source: local data frame [5 x 2]
##
## beginYear propertyDamage
## 1 2006 115993852200
## 2 2004 19591293950
## 3 2005 8250911600
## 4 1993 7604110650
## 5 1999 4605981300head(select(arrange(yearlyPropDmgSummary, desc(cropDamage)), beginYear, cropDamage),5)## Source: local data frame [5 x 2]
##
## beginYear cropDamage
## 1 1994 5590603400
## 2 1993 5421179800
## 3 2005 2596764900
## 4 2008 2010060000
## 5 1999 1930171400Conclusions
Based on this Analysis, we draw following Conclusions,
- Top Fatalities are caused by TORNADO, EXCESSIVE HEAT, HEAT, FLASH FLOOD and LIGHTNING
- Top Injuries are caused by TORNADO, THUNDERSTORM WIND, FLOOD, EXCESSIVE HEAT and LIGHTNING
- TORNADO has caused 91346 injuries compared to 9460 injuries caused by THUNDERSTORM WIND
- 5 States Worst hit by Natural Calamities are TX, MO, AL, OH and MS
- Most Disastrous Years hit by Worst Casualties are 2011, 1974, 1953, 1998 and 1965
- Top Property Damage is caused by FLOOD, HURRICANE/TYPHOON, RIVER FLOOD, TORNADO and FLASH FLOOD
- FLOOD has caused 122 Billion Dollars Property Damage compared to 37 Billion Dollars caused by HURRICANE/TYPHOON
- Top Crop Damage is caused by HURRICANE/TYPHOON, RIVER FLOOD, ICE STORM, FLOOD and HAIL
- Most Disastrous Years for Property Damage are 2006, 2004, 2005, 1993 and 1999
- Most Disastrous Years for Crop Damage are 1994, 1993, 2005, 2008 and 1999
Thanks
This completes our analysis.
Thank You for participation.