EXPLORATION OF THE U.S. NATIONAL OCIENIC AND ATMOSPHERIC ADMINISTRATION’S (NOAA) STORM DATABASE

Synopsis.

The task here is to analyse the NOAA storm database across the U.S. in order to identify the types of events that: 1. are most harmful with respect to population health? 2. have the greatest economic consequences?

From the Storm data, we will require the variables associated with population health and economic consequences in order to address the questions above.

requirements

library(knitr)
library(tidyr)
library(dplyr)

## 
## Attaching package: 'dplyr'

## The following objects are masked from 'package:stats':
## 
##     filter, lag

## The following objects are masked from 'package:base':
## 
##     intersect, setdiff, setequal, union

library(ggplot2)

Data processing

check if file exists and read it into R. If it doesnt exist, download it first then read the data

if ( file.exists("./storm data.csv.bz2")) { 
      storm = read.table("storm data.csv.bz2", header = TRUE, sep = ",", quote = "\"")
    
  } else {
  
    url = "https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2"
    download.file(url, destfile = "./storm data.csv.bz2")
    storm = read.table("storm data.csv.bz2", header = TRUE, sep = ",", quote = "\"")}

subset the columns with information on health and economic variables

storm = storm[ , c('EVTYPE', 'FATALITIES', 'INJURIES', 'PROPDMG', 'PROPDMGEXP', 
                         'CROPDMG', 'CROPDMGEXP')]

take a look

head(storm)

##    EVTYPE FATALITIES INJURIES PROPDMG PROPDMGEXP CROPDMG CROPDMGEXP
## 1 TORNADO          0       15    25.0          K       0           
## 2 TORNADO          0        0     2.5          K       0           
## 3 TORNADO          0        2    25.0          K       0           
## 4 TORNADO          0        2     2.5          K       0           
## 5 TORNADO          0        2     2.5          K       0           
## 6 TORNADO          0        6     2.5          K       0

From the storm data documentation, for damage, alphabetical characters used to signify magnitude include “K” for thousands, “M” for millions, and “B” for billions etc. We can express the PROPDMG and CROPDMG variableS in actual dollar amounts

what are the values in CROPDMGEXP

unique(storm$PROPDMGEXP)

##  [1] K M   B m + 0 5 6 ? 4 2 3 h 7 H - 1 8
## Levels:  - ? + 0 1 2 3 4 5 6 7 8 B h H K m M

PROPDMG as actual dollar amounts

for (i in 1:nrow(storm)) {
  if (storm$PROPDMGEXP [i] == "K") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^3
  } else if (storm$PROPDMGEXP [i] == "M") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^6
  }  else if (storm$PROPDMGEXP [i] == "B") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^9
  } else if (storm$PROPDMGEXP [i] == "1") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^1
  } else if (storm$PROPDMGEXP [i] == "2") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^2
  } else if (storm$PROPDMGEXP [i] == "3") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^3
  } else if (storm$PROPDMGEXP [i] == "4") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^4
  } else if (storm$PROPDMGEXP [i] == "5") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^5
  } else if (storm$PROPDMGEXP [i] == "6") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^6
  } else if (storm$PROPDMGEXP [i] == "7") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^7
  } else if (storm$PROPDMGEXP [i] == "8") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^8
  } else if (storm$PROPDMGEXP [i] == "h") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^2
  } else if (storm$PROPDMGEXP [i] == "H") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^2
  } else if (storm$PROPDMGEXP [i] == "m") {
    storm$PROPDMG [i] = storm$PROPDMG [i] *10^6
  } else {
  NULL
  }
}

what are the values in CROPDMGEXP

unique(storm$CROPDMGEXP)

## [1]   M K m B ? 0 k 2
## Levels:  ? 0 2 B k K m M

CROPDMG as actual dollar amounts

for (i in 1:nrow(storm)) {
  if (storm$CROPDMGEXP [i] == "K") {
    storm$CROPDMG [i] = storm$CROPDMG [i] *10^3
  } else if (storm$CROPDMGEXP [i] == "M") {
    storm$CROPDMG [i] = storm$CROPDMG [i] *10^6
  }  else if (storm$CROPDMGEXP [i] == "B") {
    storm$CROPDMG [i] = storm$CROPDMG [i] *10^9
  } else if (storm$CROPDMGEXP [i] == "k") {
     storm$CROPDMG [i] = storm$CROPDMG [i] *10^3
  } else if (storm$CROPDMGEXP [i] == "m") {
     storm$CROPDMG [i] = storm$CROPDMG [i] *10^6
  } else if (storm$CROPDMGEXP [i] == "2") {
     storm$CROPDMG [i] = storm$CROPDMG [i] *10^2
  } else {
    NULL
  }
}

In concidering both fatalities and injuries together as harm related to population health and property and crop damage as economic consequences, let us create a single variable that contains the sum of fatalities and injuries and another column to represent the sum of crop and property damage.

for (i in 1:nrow(storm)) {
  storm$health [i] = storm$FATALITIES [i] + storm$INJURIES [i] 
  storm$economic [i] = storm$PROPDMG [i] + storm$CROPDMG [i]
}

Now, since the two variables we just created are all we shall need to perform the analysis to answer the required questions, let us subset the dataset to have only this variables.

storm = storm[ , c('EVTYPE', 'health', 'economic')]

Since the analysis involves examining the outcomes by the type of event, let us group the data by event type variable EVTYPE

storm2 = aggregate(storm[c("health","economic")], by=list(evtype=storm$EVTYPE), FUN=sum, na.rm=TRUE)

Results.

Now let us order the data based on the event type with the highest economic impact

top_econ_impact = storm2[with(storm2, order(-economic, evtype)), ]
top_econ_impact = top_econ_impact [ , c('evtype', 'economic')]

The table below shows the event types with the 10 highest economic impact

top_econ_impact = top_econ_impact [1:10,]
top_econ_impact

##                evtype     economic
## 170             FLOOD 150319678257
## 411 HURRICANE/TYPHOON  71913712800
## 834           TORNADO  57362333947
## 670       STORM SURGE  43323541000
## 244              HAIL  18761221986
## 153       FLASH FLOOD  18243991079
## 95            DROUGHT  15018672000
## 402         HURRICANE  14610229010
## 590       RIVER FLOOD  10148404500
## 427         ICE STORM   8967041360

qplot(x=evtype, y=economic/1000000, data = top_econ_impact, stat = "identity",
                       position = "dodge")  + xlab("Event type") +
                      ylab("Total economic impact in millions (USD)") + 
                       ggtitle("10 events with highest economic impact") + theme(axis.text.x                        = element_text(angle = 90,   hjust = 1))

## Warning: `stat` is deprecated

## Warning: `position` is deprecated

based on the table and graph, we see that, floods have the most economic loss (sum of total crop and property damage). The second and that most costly events are hurricane/typhoon and tornado respectively

Now let us order the data based on the event type with the highest health impact

top_health_impact = storm2[with(storm2, order(-health, evtype)), ]
top_health_impact = top_health_impact [ , c('evtype', 'health')]

The table below shows the event types with the 10 highest health impact

top_health_impact = top_health_impact [1:10,]
top_health_impact

##                evtype health
## 834           TORNADO  96979
## 130    EXCESSIVE HEAT   8428
## 856         TSTM WIND   7461
## 170             FLOOD   7259
## 464         LIGHTNING   6046
## 275              HEAT   3037
## 153       FLASH FLOOD   2755
## 427         ICE STORM   2064
## 760 THUNDERSTORM WIND   1621
## 972      WINTER STORM   1527

qplot(x=evtype, y=health, data = top_health_impact, stat = "identity",
                       position = "dodge") +  xlab("Event type") + ylab("Total health") + 
                       ggtitle("10 events with highest health impact") + theme(axis.text.x =                        element_text(angle = 90, hjust = 1))

## Warning: `stat` is deprecated

## Warning: `position` is deprecated

based on the table and graph, we see that, tornadoes cause more health consequences (sum of total injuries and fatalities) compared to the other event types. The second and third most harmful events are excessive heat and TSTM WIND respectively.