Most harmful storm event in the US, a Coursera Assignmemt, week 4 of Reproducible Research

Synopsis

This is for an assignment of Coursera course ‘Reproducible Research’. The most harmful type of storm event against 1) population health and 2) economy was explored in U.S. National Oceanic and Atmospheric Administration’s (NOAA) storm database. The level of harms to population health was determined by a number of casualties with death or injury due to each storm event type in the country and the level of harms to economy was defined as a damage to properties, expressed in US dollar unit, for each storm event type. Against both health and economy, tornado had the greatest impact.

Data Processing

The data file repdata-data-StormData.csv.bz2 was read using the read.csv function and the variables of EVTYPE, FATALITIES, INJURIES and PROPDMG were selected out. Then, a new variable TOTAL_CASUALTIES, i.e. sum of FATALITIES and INJURIES, was created. Finally, for each storm type, the sum of TOTAL_CASUALTIES and the sum of PROPDMG were calculated then, the rank order of TOTAL_CASUALTIES and PROPDMG was used to answer the question: what type of strom event is most harmful to human health and economy. For discernbility reason, only top 5 harmful events are shown.

library(ggplot2)
storm <- read.csv("repdata-data-StormData.csv.bz2", header = T)
data <- storm[, c("EVTYPE", "FATALITIES","INJURIES", "PROPDMG")]
data$TOTAL_CASUALTIES <- data$FATALITIES + data$INJURIES
data2 <- aggregate(data[,2:5], list(data$EVTYPE), sum)
names(data2)[1] <- "EVENT_TYPE"                                         
health <- data2[order(data2$TOTAL_CASUALTIES, decreasing = T),c(1,5)]
health <- health[1:5,] # top 5 harmful EVTYPE
health$EVENT_TYPE <- factor(health$EVENT_TYPE, levels=health$EVENT_TYPE)

economy<- data2[order(data2$PROPDMG, decreasing = T),] 
economy <- economy[1:5, c(1,4)] #top 5 harmful EVTYPE to economy
economy$EVENT_TYPE <- factor(economy$EVENT_TYPE, levels=economy$EVENT_TYPE)

Results

1) The most harmful types of storm against population health

Table 1 shows 5 most harmful storm event types (EVENT_TYPE) based on the number of casualties (TOTAL_CASUALTIES; killed or injured). These data are also displayed in Figure 1. Tornado events caused the highest number of deaths or injuries across the US.

print("Table 1")

## [1] "Table 1"

print(health)

##         EVENT_TYPE TOTAL_CASUALTIES
## 834        TORNADO            96979
## 130 EXCESSIVE HEAT             8428
## 856      TSTM WIND             7461
## 170          FLOOD             7259
## 464      LIGHTNING             6046

ggplot(health, aes(EVENT_TYPE, TOTAL_CASUALTIES/10^3)) + geom_point() +
ggtitle("Fig.1 The number of death or injury from 5 most harmful storm types")+
  ylab("Number of death or injury, thousand")+
  xlab("Types of storm")

2) The most harmful types of storm against economy

Table 2 shows 5 most harmful storm types based on the property damage in US dollars. These data are also displayed in Figure 2. As is the case with human health, Tornado events had the highest impacts to economy across the US.

print("Table 2")

## [1] "Table 2"

print(economy)

##            EVENT_TYPE   PROPDMG
## 834           TORNADO 3212258.2
## 153       FLASH FLOOD 1420124.6
## 856         TSTM WIND 1335965.6
## 170             FLOOD  899938.5
## 760 THUNDERSTORM WIND  876844.2

ggplot(economy, aes(EVENT_TYPE, PROPDMG/10^3)) + geom_point() +
  ylim(0,3500) +
  ggtitle("Fig.2 Property damage from 5 most harmful storm types")+
  ylab("Property Damage in thousand US dollar")+
  xlab("Types of storm")