This research is aimed at finding the most harmful severe weather events on human health and economy. The dataset is from U.S. National Oceanic and Atmospheric Administration (NOAA). The impacts on human health is reflected as fatalities and injuries, and the impacts on economy is relfected as monetary property damage and crop damage in the dataset.The fatalities, injuries, and ecomomic damage (property damage plus crop damage) are aggregated for each weather events. Then the results are order from high to low to find out the most harmful weather events to human and economy saperately.
In this section, the data will be loaded into R with read.csv(bzfile()) function.
setwd("/NAS/jhuang/Projects/RTest/assignment2")
dt <- read.csv(bzfile("repdata-data-StormData.csv.bz2"))
The impacts of server weather events on human health are reflected in the fatalities and injuries. Thus the fatalities and injuries are aggregated by event types to calculate the effect of each event on human health. The results are then ordered from high to low to find out the most harmful events.
dt.fat <- aggregate(FATALITIES ~ EVTYPE, data = dt, sum, na.rm = T)
dt.inj <- aggregate(INJURIES ~ EVTYPE, data = dt, sum, na.rm = T)
dt.fat <- dt.fat[order(-dt.fat$FATALITIES), ]
dt.inj <- dt.inj[order(-dt.inj$INJURIES), ]
The economic consequence is represented by the damage of property and crop. The unit of property and crop damage are different in each record. For property damage, the units are contained in the column 'PROPDMGEXP', which include the following different symbols: , +, -, 0, 1, 2, 3, 4, 5, 6, 7, 8, ?, B, H, K, M, h, m, and the units for crop damage are contained in the column 'CROPDMGEXP', including the following symbols: , 0, 2, ?, B, K, M, k, m. These symbols are then converted to integer units as follows:
The new integer units are stored in the new columns named “PROPDMGUNIT” and “CROPDMGUNIT” saperately. The property damage and crop damage are then mutiplied with the corresponding damage unit column, which are then sum together and aggregated by event types to estimate their total economic impacts (stored in new column named “ECODMG”).The results are then ordered from high to low to find out the most harmful events.
dt$PROPDMGUNIT <- ifelse(dt$PROPDMGEXP == "B", 10^9, ifelse(dt$PROPDMGEXP ==
"M" | dt$PROPDMGEXP == "m", 10^6, ifelse(dt$PROPDMGEXP == "K" | dt$PROPDMGEXP ==
"k", 10^3, ifelse(dt$PROPDMGEXP == "H" | dt$PROPDMGEXP == "h", 100, 0))))
dt$CROPDMGUNIT <- ifelse(dt$CROPDMGEXP == "B", 10^9, ifelse(dt$CROPDMGEXP ==
"M" | dt$CROPDMGEXP == "m", 10^6, ifelse(dt$CROPDMGEXP == "K" | dt$CROPDMGEXP ==
"k", 10^3, 0)))
dt.econ <- aggregate(PROPDMG * PROPDMGUNIT + CROPDMG * CROPDMGUNIT ~ EVTYPE,
data = dt, sum, na.rm = T)
colnames(dt.econ) <- c("EVTYPE", "ECODMG")
dt.econ <- dt.econ[order(-dt.econ$ECODMG), ]
In the result section, the events with highest impacts on human health (fatalities and injuries) and economy will be displayed as figures.
From the following two figures, we can find that tornado is the most harmful weather events to human health (both fatalities and injuries). In terms of fatalities, excessive heat is the second most harmful weather events, followed by flash flood. However, in terms of injuries, TSTM wind and flood are the second and third most harmful weather events.
library(ggplot2)
ggplot(data = dt.fat[1:20, ], aes(x = reorder(EVTYPE, FATALITIES), y = FATALITIES)) +
geom_bar(stat = "identity") + labs(x = "Event types", y = "Fatalities",
title = "Fatalities Casused by top 20 Weather Events") + coord_flip()
ggplot(data = dt.inj[1:20, ], aes(x = reorder(EVTYPE, INJURIES), y = INJURIES)) +
geom_bar(stat = "identity") + labs(x = "Event types", y = "INJURIES", title = "Injuries Casused by top 20 Weather Events") +
coord_flip()
From the following figure, we can see that flood is the most harmful weather events to economy, followed by hurricane/typhoon, and tornado.
ggplot(data = dt.econ[1:20, ], aes(x = reorder(EVTYPE, ECODMG), y = ECODMG)) +
geom_bar(stat = "identity") + labs(x = "Event types", y = "Economic Damage (dollars)",
title = "Economic Damage Casused by top 20 Weather Events") + coord_flip()
