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.
This project involves exploring the U.S. National Oceanic and Atmospheric Administration’s (NOAA) storm database. This database 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.
Loading libraries for data reading.
## Libraries
library(dplyr)
library(ggplot2)
library(knitr)
Loading data from zip file in the current directory. If it is not found, it is downloaded by the indicated URL.
## Download Data
filename1 <- "./repdata_data_StormData.csv.bz2"
if(!file.exists(filename1)){
#zipname <- "repdata_data_activity.zip"
fileURL <- "https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2"
download.file(fileURL, destfile = filename1, method = "curl")
}
rawdata <- read.csv(filename1)
tidydata <- select(rawdata,EVTYPE,FATALITIES:CROPDMGEXP) #Resumed data for analyses.
NAvalues <- sum(is.na(tidydata))
In the data there is 0 NA value in the data.
Processing data for analysis of most harmful eventy type with respect to population health.
tidydata <- group_by(tidydata,EVTYPE)
## Fatalities Processing Data
US_FTL <- summarise(tidydata,FTL = sum(FATALITIES))
US_FTL <- US_FTL %>% arrange(desc(FTL)) %>% mutate(FTLPER = FTL/sum(FTL)*100,FTLCUM = cumsum(FTLPER))
#Injuries Processing Data
US_INJ <- summarise(tidydata,INJ = sum(INJURIES))
US_INJ <- US_INJ %>% arrange(desc(INJ)) %>% mutate(INJPER = INJ/sum(INJ)*100,INJCUM = cumsum(INJPER))
Processing data for analysis of most harmful eventy type with respect of economic consequences. The factors used on PROPDMGEXP and CROPDMGEXP were correlated with multipliers described in this link.
# Economic Processing Data
tidydata <- ungroup(tidydata)
Econ_data <- select(tidydata,EVTYPE,PROPDMG,PROPDMGEXP,CROPDMG,CROPDMGEXP) # Economic Impact Data
Prop_factors <- data.frame(Factor = levels(Econ_data$PROPDMGEXP),Multiplier = c(rep(0,3),10^0,rep(10^1,9),10^9,10^2,10^2,10^3,10^6,10^6))
Crop_factors <- data.frame(Factor = levels(Econ_data$CROPDMGEXP),Multiplier = c(rep(0,2),rep(10^1,2),10^9,10^3,10^3,10^6,10^6))
Econ_data <- Econ_data %>% mutate(PROPDMG_T = PROPDMG*Prop_factors$Multiplier[match(Econ_data$PROPDMGEXP,Prop_factors$Factor)],CROPDMG_T = CROPDMG*Crop_factors$Multiplier[match(Econ_data$CROPDMGEXP,Crop_factors$Factor)],ECONDMG = PROPDMG_T + CROPDMG_T) %>% group_by(EVTYPE) # Added columns with total amount of properties damage (PROPDMG_T), total amount of crop damage (CROPDMG_T), and total amount of economic impact
Econ_Evtype <- summarise(Econ_data,EVTYPEDMG = sum(ECONDMG))
Econ_Evtype <- Econ_Evtype %>% arrange(desc(EVTYPEDMG)) %>% mutate(DMGPER = EVTYPEDMG/sum(EVTYPEDMG)*100) %>% mutate(DMGCUM = cumsum(DMGPER))
ggplot(US_FTL[1:10,],aes(x=reorder(EVTYPE,FTLCUM)))+geom_col(aes(y=FTL))+labs(x = "Event Type", y = "Fatalities", title = "Fatalities in US by Type of Event")+theme_bw(base_size = 7)+scale_y_continuous(sec.axis = sec_axis(~./55, name = "Cumulative Fatalities in US [%]"))+geom_point(aes(y=FTLCUM*55),pch = 16, size = 2, color = "darkblue")+geom_path(aes(y=FTLCUM*55, group=1), colour="darkblue", lty=3, size=0.9)
The highest fatalities by event type. The parameters are:
head(US_FTL,10)
## # A tibble: 10 x 4
## EVTYPE FTL FTLPER FTLCUM
## <fct> <dbl> <dbl> <dbl>
## 1 TORNADO 4179 57.9 57.9
## 2 HEAT 701 9.71 67.6
## 3 TSTM WIND 295 4.09 71.7
## 4 LIGHTNING 241 3.34 75.0
## 5 FLASH FLOOD 226 3.13 78.1
## 6 EXCESSIVE HEAT 179 2.48 80.6
## 7 HEAT WAVE 172 2.38 83.0
## 8 EXTREME COLD 103 1.43 84.4
## 9 FLOOD 103 1.43 85.9
## 10 EXTREME HEAT 96 1.33 87.2
ggplot(US_INJ[1:10,],aes(x= reorder(EVTYPE,INJCUM)))+geom_col(aes(y=INJ))+labs(x = "Event Type", y = "Injuries", title = "Injuries in US by Type of Event")+theme_bw(base_size = 7)+scale_y_continuous(sec.axis = sec_axis(~./1000, name = "Cumulative Injuries in US [%]"))+geom_point(aes(y=INJCUM*1000),pch = 16, size = 2, color = "darkblue")+geom_path(aes(y=INJCUM*1000, group=1), colour="darkblue", lty=3, size=0.9)
The highest injuries by event type. The parameters are:
head(US_INJ,10)
## # A tibble: 10 x 4
## EVTYPE INJ INJPER INJCUM
## <fct> <dbl> <dbl> <dbl>
## 1 TORNADO 72055 82.7 82.7
## 2 TSTM WIND 3808 4.37 87.1
## 3 ICE STORM 1800 2.07 89.2
## 4 LIGHTNING 1578 1.81 91.0
## 5 THUNDERSTORM WINDS 908 1.04 92.0
## 6 HEAT 878 1.01 93.0
## 7 HAIL 730 0.838 93.9
## 8 BLIZZARD 596 0.684 94.6
## 9 EXCESSIVE HEAT 546 0.627 95.2
## 10 HEAVY SNOW 486 0.558 95.8
ggplot(Econ_Evtype[1:10,],aes(x=reorder(EVTYPE,DMGCUM)))+geom_col(aes(y=EVTYPEDMG/10^6))+labs(x = "Event Type", y = "Economic Damage [MiUSD]", title = "Economic Damage by Type of Event")+theme_bw(base_size = 7)+scale_y_continuous(sec.axis = sec_axis(~./350, name = "Cumulative Economic Damage in US [%]"))+geom_point(aes(y=DMGCUM*350),pch = 16, size = 2, color = "darkblue")+geom_path(aes(y=DMGCUM*350, group=1), colour="darkblue", lty=3, size=0.9)
The highest economic impacts by event type. The parameters are:
head(Econ_Evtype,10)
## # A tibble: 10 x 4
## EVTYPE EVTYPEDMG DMGPER DMGCUM
## <fct> <dbl> <dbl> <dbl>
## 1 TORNADO 33556245357 39.1 39.1
## 2 RIVER FLOOD 10133367500 11.8 50.9
## 3 ICE STORM 5416705150 6.31 57.2
## 4 WINTER STORM 5217677050 6.07 63.2
## 5 FLOOD 3886654300 4.53 67.8
## 6 HURRICANE OPAL 3191846000 3.72 71.5
## 7 FLASH FLOOD 2564411401 2.99 74.5
## 8 HAIL 2549130097 2.97 77.4
## 9 HEAVY RAIN/SEVERE WEATHER 2500000000 2.91 80.4
## 10 THUNDERSTORM WINDS 1926669872 2.24 82.6