Storm Analysis with NOOA Storm Database From 1950 - 2011

Synopsis

The goal of this project is to explore the U.S. National Oceanic and Atmospheric Administration’s (NOOA) Storm Database and answer some basic questions about severe weather events, for example:

• Which types of events are most harmful to population health ?
• Which types of events have the greatest economic consequences ?

The database tracks charateristics of major storms and weather events in the United States from 1950 to 2011, including when and where they occur, as well as estimates of any fatalities, injures, and property damage.

Data and Documentation

• Storm Data [47Mb]
• National Weather Service Storm Data Documentation
• National Climatic Data Center Storm Events FAQ

Data Processing

• Data were downloaded directly and read into data frame for processing.
• For the purpose of evaluating damage/fatilities/injuries of storm events, I removed all events with 0 damage/deaths/injury and subtract only 4 important columns: FATALITIES, INJURIES, PROPDMG, and CROPDMG
• To estimate the impact of events to population health, I calculate total number of fatalities and injuries for each event type, based on two columns FATALITIES and INJURIES, and extract top 10 harmful events
• To estimate the impact of events to economy, I calculate total damage in property and crop, based on two columns PROPDMG and CROPDMG, then subset top 10 biggest damage

• Other cleaning task: Leading and trailing white space were removed in the Event type names; upper cases also were converted into lower case.

• Load data

download.file("https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2", "stormData.csv.bz2", method = "curl")
storm <- read.table("stormData.csv.bz2", header = T, sep = ",", stringsAsFactors = F)

• Subset important columns, remove 0 damage/injury/fatality

stormReduce <- subset(storm, select = c(EVTYPE, FATALITIES, INJURIES, PROPDMG, CROPDMG))
stormReduce <- subset(stormReduce,PROPDMG != 0 | CROPDMG != 0 | FATALITIES != 0 | INJURIES !=0)

• Clean Data in EVTYPE column

library(stringr)
stormReduce$EVTYPE <- tolower(str_trim(stormReduce$EVTYPE))

• Calculate event’s damage to population health

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, union

storm_HealthDamage <- stormReduce %>% group_by(EVTYPE) %>% summarise(FATALITIES = sum(FATALITIES), INJURIES = sum(INJURIES))
# Calculate top 10 damages 
storm_HealthDamage$DAMAGE <- storm_HealthDamage$FATALITIES + storm_HealthDamage$INJURIES
top10HealthDamage <- head(storm_HealthDamage[order(storm_HealthDamage$DAMAGE, decreasing = T),],10)
top10HealthDamage$EVTYPE <- as.factor(top10HealthDamage$EVTYPE)
# Reorder factor based on value of damage 
top10HealthDamage$EVTYPE <- reorder(top10HealthDamage$EVTYPE, top10HealthDamage$DAMAGE, sum)
top10HealthDamage

## Source: local data frame [10 x 4]
## 
##               EVTYPE FATALITIES INJURIES DAMAGE
## 1            tornado       5633    91346  96979
## 2     excessive heat       1903     6525   8428
## 3          tstm wind        504     6957   7461
## 4              flood        470     6789   7259
## 5          lightning        816     5230   6046
## 6               heat        937     2100   3037
## 7        flash flood        978     1777   2755
## 8          ice storm         89     1975   2064
## 9  thunderstorm wind        133     1488   1621
## 10      winter storm        206     1321   1527

• Calculate event’s damage to economics consequences

storm_EcoDamage <- stormReduce %>% group_by(EVTYPE) %>% summarise(PROPDMG = sum(PROPDMG), CROPDMG = sum(CROPDMG))
storm_EcoDamage$DAMAGE <- storm_EcoDamage$PROPDMG + storm_EcoDamage$CROPDMG
top10EcoDamage <- head(storm_EcoDamage[order(storm_EcoDamage$DAMAGE, decreasing = T),], 10)
top10EcoDamage$EVTYPE <-  factor(top10EcoDamage$EVTYPE)
top10EcoDamage$EVTYPE <- reorder(top10EcoDamage$EVTYPE, top10EcoDamage$DAMAGE, sum)
top10EcoDamage

## Source: local data frame [10 x 4]
## 
##                EVTYPE   PROPDMG   CROPDMG    DAMAGE
## 1             tornado 3212258.2 100018.52 3312276.7
## 2         flash flood 1420174.6 179200.46 1599375.1
## 3           tstm wind 1336103.6 109202.60 1445306.2
## 4                hail  688693.4 579596.28 1268289.7
## 5               flood  899938.5 168037.88 1067976.4
## 6   thunderstorm wind  876844.2  66791.45  943635.6
## 7           lightning  603351.8   3580.61  606932.4
## 8  thunderstorm winds  446293.2  18684.93  464978.1
## 9           high wind  324731.6  17283.21  342014.8
## 10       winter storm  132720.6   1978.99  134699.6

• Reshape data frame for visualization

library(reshape2)
top10HealthDamage <- melt(top10HealthDamage[,1:3], id.vars = "EVTYPE", variable.name = "DAMAGE_TYPE", value.name = "DAMAGE_COUNT")
top10EcoDamage <- melt(top10EcoDamage[,1:3], id.vars = "EVTYPE", variable.name = "DAMAGE_TYPE", value.name = "DAMAGE_COUNT")
# Change to descriptive label
top10EcoDamage$DAMAGE_TYPE <- factor(top10EcoDamage$DAMAGE_TYPE, labels = c("PROPERTY_DAMAGE", "CROP_DAMAGE"))

Result

Top 10 Harmful Events to Population Health

library(ggplot2)
ggplot(top10HealthDamage, aes(EVTYPE, DAMAGE_COUNT, fill = DAMAGE_TYPE)) + geom_bar(stat = "identity", position = "dodge") + coord_flip()

Top 10 Harmful Events to Economy

ggplot(top10EcoDamage, aes(EVTYPE, DAMAGE_COUNT, fill = DAMAGE_TYPE)) + geom_bar(stat = "identity", position = "dodge") + coord_flip()