Reproducible Research Course Project 2: Top 10 Severe weather events in the USA from 1995-2011 according to health and economic impacts
Alex Hall
11/01/2022
Synopsis
Overview
This analysis investigates the following questions:
- Across the United States, which types of events (as indicated in the EVTYPE variable) are most harmful with respect to population health?
- Across the United States, which types of events have the greatest economic consequences?
Excessive heat appears to be the most harmful event in terms of fatalities per single event (mean) and across all recorded events (total). In terms of injuries, excessive heat again causes the most injuries per event (mean), although tsunamis are not far behind. However, tornadoes come out on top in terms of total injuries.
Hurricanes/typhoons are the most destructive in terms of property damage, per single event, but floods have caused more damage over the recorded events in this dataset. Droughts are by far the most destructive for crops, per single event and across all events.
Limitations
As described in the data processing section, accuracy compromises were made in favour of efficiency and this also extends to fully exploring each events’ remarks and any potential outliers. Also large ranges were observed in the health and economic impact values and simple statistics such as mean and total probably do not fully account for the intricacies in the data and things such as likelihood of more severe events - Further exploratory analysis is recommended.
Load packages
library(tidyverse)
library(knitr)
library(lubridate)
library(fuzzyjoin)
library(stringdist)
library(gridExtra)##
## Attaching package: 'gridExtra'## The following object is masked from 'package:dplyr':
##
## combineData Processing
This next section details the steps that were taken to source the data and process it in preparation for analysis. Broadly speaking, the steps were:
- Download and read data
- Remove the variables that were not required to answer the key questions
- Clean the date variable and remove records prior to Jan 1996
- Remove records with no injuries, property or crop damage
- Clean (standarise) the property and crop damage variables
- Clean the event type variable “EVTYPE”
Download and read data
This first code chunk downloads and reads in the raw data before displaying some summary information.
## Check if data has already been downloaded
# Create object with name of destination file
destfile <- "StormData.csv"
# Run check. If the file DOES NOT exist, then it downloads it.
if (!file.exists(destfile))
{
download.file("https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2",
destfile = destfile)
}
## Read data into R object
StormDataRaw <- read.csv("StormData.csv")
## View quick summary of raw data
print(str(StormDataRaw))## 'data.frame': 902297 obs. of 37 variables:
## $ STATE__ : num 1 1 1 1 1 1 1 1 1 1 ...
## $ BGN_DATE : chr "4/18/1950 0:00:00" "4/18/1950 0:00:00" "2/20/1951 0:00:00" "6/8/1951 0:00:00" ...
## $ BGN_TIME : chr "0130" "0145" "1600" "0900" ...
## $ TIME_ZONE : chr "CST" "CST" "CST" "CST" ...
## $ COUNTY : num 97 3 57 89 43 77 9 123 125 57 ...
## $ COUNTYNAME: chr "MOBILE" "BALDWIN" "FAYETTE" "MADISON" ...
## $ STATE : chr "AL" "AL" "AL" "AL" ...
## $ EVTYPE : chr "TORNADO" "TORNADO" "TORNADO" "TORNADO" ...
## $ BGN_RANGE : num 0 0 0 0 0 0 0 0 0 0 ...
## $ BGN_AZI : chr "" "" "" "" ...
## $ BGN_LOCATI: chr "" "" "" "" ...
## $ END_DATE : chr "" "" "" "" ...
## $ END_TIME : chr "" "" "" "" ...
## $ COUNTY_END: num 0 0 0 0 0 0 0 0 0 0 ...
## $ COUNTYENDN: logi NA NA NA NA NA NA ...
## $ END_RANGE : num 0 0 0 0 0 0 0 0 0 0 ...
## $ END_AZI : chr "" "" "" "" ...
## $ END_LOCATI: chr "" "" "" "" ...
## $ LENGTH : num 14 2 0.1 0 0 1.5 1.5 0 3.3 2.3 ...
## $ WIDTH : num 100 150 123 100 150 177 33 33 100 100 ...
## $ F : int 3 2 2 2 2 2 2 1 3 3 ...
## $ MAG : num 0 0 0 0 0 0 0 0 0 0 ...
## $ FATALITIES: num 0 0 0 0 0 0 0 0 1 0 ...
## $ INJURIES : num 15 0 2 2 2 6 1 0 14 0 ...
## $ PROPDMG : num 25 2.5 25 2.5 2.5 2.5 2.5 2.5 25 25 ...
## $ PROPDMGEXP: chr "K" "K" "K" "K" ...
## $ CROPDMG : num 0 0 0 0 0 0 0 0 0 0 ...
## $ CROPDMGEXP: chr "" "" "" "" ...
## $ WFO : chr "" "" "" "" ...
## $ STATEOFFIC: chr "" "" "" "" ...
## $ ZONENAMES : chr "" "" "" "" ...
## $ LATITUDE : num 3040 3042 3340 3458 3412 ...
## $ LONGITUDE : num 8812 8755 8742 8626 8642 ...
## $ LATITUDE_E: num 3051 0 0 0 0 ...
## $ LONGITUDE_: num 8806 0 0 0 0 ...
## $ REMARKS : chr "" "" "" "" ...
## $ REFNUM : num 1 2 3 4 5 6 7 8 9 10 ...
## NULLprint(head(StormDataRaw))## STATE__ BGN_DATE BGN_TIME TIME_ZONE COUNTY COUNTYNAME STATE EVTYPE
## 1 1 4/18/1950 0:00:00 0130 CST 97 MOBILE AL TORNADO
## 2 1 4/18/1950 0:00:00 0145 CST 3 BALDWIN AL TORNADO
## 3 1 2/20/1951 0:00:00 1600 CST 57 FAYETTE AL TORNADO
## 4 1 6/8/1951 0:00:00 0900 CST 89 MADISON AL TORNADO
## 5 1 11/15/1951 0:00:00 1500 CST 43 CULLMAN AL TORNADO
## 6 1 11/15/1951 0:00:00 2000 CST 77 LAUDERDALE AL TORNADO
## BGN_RANGE BGN_AZI BGN_LOCATI END_DATE END_TIME COUNTY_END COUNTYENDN
## 1 0 0 NA
## 2 0 0 NA
## 3 0 0 NA
## 4 0 0 NA
## 5 0 0 NA
## 6 0 0 NA
## END_RANGE END_AZI END_LOCATI LENGTH WIDTH F MAG FATALITIES INJURIES PROPDMG
## 1 0 14.0 100 3 0 0 15 25.0
## 2 0 2.0 150 2 0 0 0 2.5
## 3 0 0.1 123 2 0 0 2 25.0
## 4 0 0.0 100 2 0 0 2 2.5
## 5 0 0.0 150 2 0 0 2 2.5
## 6 0 1.5 177 2 0 0 6 2.5
## PROPDMGEXP CROPDMG CROPDMGEXP WFO STATEOFFIC ZONENAMES LATITUDE LONGITUDE
## 1 K 0 3040 8812
## 2 K 0 3042 8755
## 3 K 0 3340 8742
## 4 K 0 3458 8626
## 5 K 0 3412 8642
## 6 K 0 3450 8748
## LATITUDE_E LONGITUDE_ REMARKS REFNUM
## 1 3051 8806 1
## 2 0 0 2
## 3 0 0 3
## 4 0 0 4
## 5 0 0 5
## 6 0 0 6Remove the variables that were not required to answer the key questions
In order to reduce the size of the dataset, variables that were not required to answer the key questions were removed.
## Subset raw data to only include variables needed for analysis to answer questions of interest
StormData <- StormDataRaw %>%
select(BGN_DATE,
EVTYPE,
FATALITIES,
INJURIES,
PROPDMG,
PROPDMGEXP,
CROPDMG,
CROPDMGEXP,
REFNUM)Clean the date variable and remove records prior to Jan 1996
The original date variable was formatted as a date variable. Then records prior to Jan 1996 were removed. This is because this dataset only recorded tornadoes up until then which heavily skewed the data towards tornadoes. It was deemed preferable to include years in which all storm events were recorded in order to compare them.
## Convert BGN_DATE into BGN_DATE_CLEAN variable, formatted as a date.
StormData <-
StormData %>%
mutate(BGN_DATE_CLEAN =
mdy(
str_extract(
StormData$BGN_DATE,"[:graph:]*(?=[:space:])"
)
)
)
## Drop old date variable
StormData <- select(StormData, -BGN_DATE)
## Subset to records from Jan 1996 when all weather events are recorded
StormData <- filter(StormData, BGN_DATE_CLEAN >= ymd("1995/01/01"))Remove records with no injuries, property or crop damage
To further reduce the dataset, records with no recorded harmful consequences were removed.
## Subset to remove records with 0 fatalities, injuries, crop or property damage
StormData <-
StormData %>%
filter((FATALITIES + INJURIES + PROPDMG + CROPDMG) != 0
)Clean (standarise) the property and crop damage variables
The original dataset included the value and scale as separate variables, so these were combined.
## Change exp variables to lower case so that B and b are the same
StormData$PROPDMGEXP <- tolower(StormData$PROPDMGEXP)
StormData$CROPDMGEXP <- tolower(StormData$CROPDMGEXP)
## Convert DMG and EXP variables into single variable with numeric value
StormData <-
StormData %>%
mutate(PROPDMG_CLEAN =
case_when(
PROPDMGEXP == "b" ~ PROPDMG * 10^9,
PROPDMGEXP == "m" | PROPDMGEXP == "m" ~ PROPDMG * 10^6,
PROPDMGEXP == "k" ~ PROPDMG * 10^3,
PROPDMGEXP == "h" ~ PROPDMG * 10^2,
PROPDMGEXP %in%
c("0","1","2","3","4","5","6","7","8","9")
~ PROPDMG * 10^1,
PROPDMGEXP == "+" ~ PROPDMG,
PROPDMGEXP == "-" ~ NA_real_,
PROPDMGEXP == "?" | PROPDMGEXP == "" ~ PROPDMG * 0,
TRUE ~ NA_real_
),
CROPDMG_CLEAN =
case_when(
CROPDMGEXP == "b" ~ CROPDMG * 10^9,
CROPDMGEXP == "m" | CROPDMGEXP == "m" ~ CROPDMG * 10^6,
CROPDMGEXP == "k" ~ CROPDMG * 10^3,
CROPDMGEXP == "h" ~ CROPDMG * 10^2,
CROPDMGEXP %in%
c("0","1","2","3","4","5","6","7","8","9")
~ CROPDMG * 10^1,
CROPDMGEXP == "+" ~ CROPDMG,
CROPDMGEXP == "-" ~ NA_real_,
CROPDMGEXP == "?" | CROPDMGEXP == "" ~ CROPDMG * 0,
TRUE ~ NA_real_
)
)
## Drop original PROP and CROP dmg variables
StormData <-
select(StormData, -PROPDMG, -PROPDMGEXP, -CROPDMG, -CROPDMGEXP)Clean the event type variable “EVTYPE”
Even after the cleaning and dataset reduction conducted thus far, there were numerous errors and inconsistencies within this EVTYPE variable. A number of steps were taken to resolve this to a certain extent but it by no means a complete solution and there will therefore be some event types incorrectly assigned. However, exploratory data analysis suggests that the largest events in terms of fatalities, injuries and damage are accounted for so a more accurate assignment should not change the findings significantly.
The first code chunk loads the official list of storm event types, as transcribed from this url: https://d396qusza40orc.cloudfront.net/repdata%2Fpeer2_doc%2Fpd01016005curr.pdf It then stores it as a dataframe that can be joined with the main data.
## Load official list of storm event types for reference
StormEvents <-
as.character(
c(
"Astronomical Low Tide",
"Avalanche",
"Blizzard",
"Coastal Flood",
"Cold/Wind Chill",
"Debris Flow",
"Dense Fog",
"Dense Smoke",
"Drought",
"Dust Devil",
"Dust Storm",
"Excessive Heat",
"Extreme Cold/Wind Chill",
"Flash Flood",
"Flood",
"Frost/Freeze",
"Funnel Cloud",
"Freezing Fog",
"Hail",
"Heat",
"Heavy Rain",
"Heavy Snow",
"High Surf",
"High Wind",
"Hurricane (Typhoon)",
"Ice Storm",
"Lake-Effect Snow",
"Lakeshore Flood",
"Lightning",
"Marine Hail",
"Marine High Wind",
"Marine Strong Wind",
"Marine Thunderstorm Wind",
"Rip Current",
"Seiche",
"Sleet",
"Storm Surge/Tide",
"Strong Wind",
"Thunderstorm Wind",
"Tornado",
"Tropical Depression",
"Tropical Storm",
"Tsunami",
"Volcanic Ash",
"Waterspout",
"Wildfire",
"Winter Storm",
"Winter Weather"
)
)
## Create Storm Events dataframe
StormEventsDataFrame <- as.data.frame(StormEvents)
## Change variable name
StormEventsDataFrame <-
rename(StormEventsDataFrame, EVTYPE = StormEvents)
## Print list
print(StormEventsDataFrame)## EVTYPE
## 1 Astronomical Low Tide
## 2 Avalanche
## 3 Blizzard
## 4 Coastal Flood
## 5 Cold/Wind Chill
## 6 Debris Flow
## 7 Dense Fog
## 8 Dense Smoke
## 9 Drought
## 10 Dust Devil
## 11 Dust Storm
## 12 Excessive Heat
## 13 Extreme Cold/Wind Chill
## 14 Flash Flood
## 15 Flood
## 16 Frost/Freeze
## 17 Funnel Cloud
## 18 Freezing Fog
## 19 Hail
## 20 Heat
## 21 Heavy Rain
## 22 Heavy Snow
## 23 High Surf
## 24 High Wind
## 25 Hurricane (Typhoon)
## 26 Ice Storm
## 27 Lake-Effect Snow
## 28 Lakeshore Flood
## 29 Lightning
## 30 Marine Hail
## 31 Marine High Wind
## 32 Marine Strong Wind
## 33 Marine Thunderstorm Wind
## 34 Rip Current
## 35 Seiche
## 36 Sleet
## 37 Storm Surge/Tide
## 38 Strong Wind
## 39 Thunderstorm Wind
## 40 Tornado
## 41 Tropical Depression
## 42 Tropical Storm
## 43 Tsunami
## 44 Volcanic Ash
## 45 Waterspout
## 46 Wildfire
## 47 Winter Storm
## 48 Winter WeatherThis chunk first does some initial cleaning of the text in the EVTYPE variable to remove leading and trailing whitespace, as well as double spaces. It also sets the text to title case to match the official list above.
It then investigates the unique values of EVTYPE, the frequencies against these unique values and the fatalities, injuries and damage associated with each. This helps to identify targets for further cleaning. E.g. which have the highest frequency or damage and therefore should be most pertinent to the analysis. This is under the assumption that it is not feasible as part of this project to clean every single value.
## Initial clean up EVTYPE
## Set EVTYPE to title case and trim whitespace
StormData$EVTYPE <- str_to_title(trimws(StormData$EVTYPE))
## Convert double space to single space
StormData$EVTYPE <- gsub("\\s+", " ", StormData$EVTYPE)
## Initial pass to identify EVTYPE value inconsistencies that might not be picked up
## by stringdist.
## Count unique values of EVTYPE
length(unique(StormData$EVTYPE))## [1] 326## Find unique values of EVTYPE
EVTYPEunique <- unique(StormData$EVTYPE)
## Check summary of EVTYPE frequencies
table(StormData$EVTYPE)##
## Agricultural Freeze Astronomical High Tide
## 3 8
## Astronomical Low Tide Avalanche
## 2 266
## Beach Erosion Black Ice
## 1 1
## Blizzard Blowing Dust
## 230 1
## Blowing Snow Breakup Flooding
## 2 1
## Brush Fire Coastal Erosion
## 1 1
## Coastal Flood Coastal Flooding
## 156 36
## Coastal Flooding/Erosion Coastal Storm
## 4 4
## Coastalstorm Cold
## 1 28
## Cold And Snow Cold And Wet Conditions
## 1 1
## Cold Temperature Cold Wave
## 2 1
## Cold Weather Cold/Wind Chill
## 4 90
## Dam Break Damaging Freeze
## 2 6
## Dense Fog Dense Smoke
## 64 1
## Downburst Drought
## 1 264
## Drought/Excessive Heat Drowning
## 10 1
## Dry Microburst Dust Devil
## 77 90
## Dust Devil Waterspout Dust Storm
## 1 98
## Early Frost Erosion/Cstl Flood
## 1 2
## Excessive Heat Excessive Rainfall
## 698 1
## Excessive Snow Excessive Wetness
## 25 1
## Extended Cold Extreme Cold
## 1 179
## Extreme Cold/Wind Chill Extreme Heat
## 111 14
## Extreme Wind Chill Extreme Windchill
## 1 19
## Falling Snow/Ice Flash Flood
## 2 19851
## Flash Flood - Heavy Rain Flash Flood Winds
## 2 1
## Flash Flood/ Street Flash Flood/Flood
## 1 3
## Flash Flooding Flash Floods
## 100 8
## Flood Flood & Heavy Rain
## 9712 2
## Flood/Flash Flood/Flash Flood
## 1 88
## Flood/Flash/Flood Flood/Rain/Winds
## 1 5
## Flooding Flooding/Heavy Rain
## 23 1
## Floods Fog
## 1 104
## Freeze Freezing Drizzle
## 16 7
## Freezing Fog Freezing Rain
## 7 18
## Freezing Rain/Snow Freezing Spray
## 3 1
## Frost Frost/Freeze
## 2 117
## Funnel Cloud Glaze
## 9 20
## Glaze Ice Gradient Wind
## 2 6
## Grass Fires Gustnado
## 1 2
## Gusty Wind Gusty Wind/Hail
## 13 1
## Gusty Wind/Hvy Rain Gusty Wind/Rain
## 1 1
## Gusty Winds Hail
## 30 23860
## Hail 0.75 Hail 075
## 1 1
## Hail 100 Hail 125
## 1 1
## Hail 150 Hail 175
## 1 2
## Hail 200 Hail 275
## 1 3
## Hail 450 Hail 75
## 1 1
## Hail Damage Hailstorm
## 1 3
## Hard Freeze Hazardous Surf
## 2 1
## Heat Heat Wave
## 203 33
## Heat Wave Drought Heavy Mix
## 1 2
## Heavy Rain Heavy Rain And Flood
## 1080 1
## Heavy Rain/High Surf Heavy Rain/Severe Weather
## 1 1
## Heavy Rains Heavy Seas
## 4 2
## Heavy Snow Heavy Snow-Squalls
## 1134 15
## Heavy Snow And High Winds Heavy Snow And Strong Winds
## 1 1
## Heavy Snow Shower Heavy Snow Squalls
## 1 12
## Heavy Snow/High Winds & Flood Heavy Snow/Ice
## 1 1
## Heavy Surf Heavy Surf And Wind
## 30 1
## Heavy Surf Coastal Flooding Heavy Surf/High Surf
## 1 50
## Heavy Swells High Seas
## 1 6
## High Surf High Surf Advisory
## 132 1
## High Swells High Water
## 2 2
## High Waves High Wind
## 1 5416
## High Wind (G40) High Wind Damage
## 2 2
## High Winds High Winds Heavy Rains
## 312 1
## Hurricane Hurricane-Generated Swells
## 126 3
## Hurricane Edouard Hurricane Erin
## 1 7
## Hurricane Felix Hurricane Opal
## 1 8
## Hurricane Opal/High Winds Hurricane/Typhoon
## 1 72
## Hvy Rain Hyperthermia/Exposure
## 1 1
## Hypothermia/Exposure Ice
## 6 2
## Ice Jam Flood (Minor Ice Jam Flooding
## 1 2
## Ice On Road Ice Roads
## 1 1
## Ice Storm Ice/Strong Winds
## 653 1
## Icy Roads Lake-Effect Snow
## 22 194
## Lake Effect Snow Lake Flood
## 4 1
## Lakeshore Flood Landslide
## 5 193
## Landslides Landslump
## 1 1
## Landspout Late Season Snow
## 2 1
## Light Freezing Rain Light Snow
## 22 141
## Light Snowfall Lightning
## 1 12026
## Lightning And Heavy Rain Lightning And Thunderstorm Win
## 1 1
## Lightning Fire Lightning Injury
## 1 1
## Lightning Thunderstorm Winds Lightning Wauseon
## 1 1
## Lightning. Ligntning
## 1 1
## Marine Accident Marine Hail
## 1 2
## Marine High Wind Marine Mishap
## 19 2
## Marine Strong Wind Marine Thunderstorm Wind
## 46 33
## Marine Tstm Wind Microburst
## 109 1
## Mixed Precip Mixed Precipitation
## 6 18
## Mud Slide Mud Slides Urban Flooding
## 2 1
## Mudslide Mudslides
## 5 1
## Non-Severe Wind Damage Non-Tstm Wind
## 1 1
## Non Tstm Wind Other
## 1 34
## Rain Rain/Snow
## 3 2
## Rapidly Rising Water Record Cold
## 1 1
## Record Heat Rip Current
## 1 389
## Rip Currents Rip Currents/Heavy Surf
## 241 2
## River And Stream Flood River Flood
## 2 95
## River Flooding Rock Slide
## 14 1
## Rogue Wave Rough Seas
## 1 3
## Rough Surf Rural Flood
## 2 1
## Seiche Severe Thunderstorm
## 9 5
## Severe Thunderstorm Winds Severe Thunderstorms
## 3 2
## Sleet Small Hail
## 1 11
## Snow Snow And Ice
## 31 3
## Snow Freezing Rain Snow Squall
## 10 3
## Snow Squalls Snow/Ice
## 10 1
## Snow/Sleet/Freezing Rain Storm Force Winds
## 2 1
## Storm Surge Storm Surge/Tide
## 169 47
## Strong Wind Strong Winds
## 3370 48
## Thundeerstorm Winds Thunderestorm Winds
## 1 1
## Thundersnow Thunderstorm
## 1 14
## Thunderstorm Damage To Thunderstorm Wind
## 1 43170
## Thunderstorm Wind (G40) Thunderstorm Wind 60 Mph
## 1 2
## Thunderstorm Wind 65 Mph Thunderstorm Wind 65mph
## 1 1
## Thunderstorm Wind 98 Mph Thunderstorm Wind G52
## 1 1
## Thunderstorm Wind G55 Thunderstorm Wind G60
## 1 1
## Thunderstorm Wind Trees Thunderstorm Wind.
## 1 1
## Thunderstorm Wind/ Tree Thunderstorm Wind/ Trees
## 1 4
## Thunderstorm Wind/Awning Thunderstorm Wind/Lightning
## 1 1
## Thunderstorm Winds Thunderstorm Winds 63 Mph
## 5098 1
## Thunderstorm Winds And Thunderstorm Winds G60
## 1 1
## Thunderstorm Winds Hail Thunderstorm Winds Lightning
## 40 6
## Thunderstorm Winds/ Flood Thunderstorm Winds/Hail
## 2 3
## Thunderstorm Winds53 Thunderstorm Windshail
## 1 1
## Thunderstorm Windss Thunderstorm Wins
## 6 1
## Thunderstorms Wind Thunderstorms Winds
## 1 4
## Thunderstormw Thundertorm Winds
## 1 3
## Thunerstorm Winds Tidal Flooding
## 1 4
## Tornado Tornado F0
## 13000 5
## Tornado F1 Tornado F2
## 4 2
## Tornado F3 Torrential Rainfall
## 2 1
## Tropical Depression Tropical Storm
## 35 412
## Tropical Storm Dean Tropical Storm Jerry
## 1 2
## Tstm Wind Tstm Wind (41)
## 61903 1
## Tstm Wind (G35) Tstm Wind (G40)
## 1 9
## Tstm Wind (G45) Tstm Wind 40
## 38 1
## Tstm Wind 45 Tstm Wind 55
## 1 2
## Tstm Wind 65) Tstm Wind And Lightning
## 1 1
## Tstm Wind Damage Tstm Wind G45
## 1 1
## Tstm Wind G58 Tstm Wind/Hail
## 1 441
## Tstm Winds Tsunami
## 5 14
## Tunderstorm Wind Typhoon
## 1 9
## Unseasonable Cold Unseasonably Cold
## 1 4
## Unseasonably Warm Unseasonably Warm And Dry
## 7 1
## Unseasonal Rain Urban And Small Stream Floodin
## 2 1
## Urban Flood Urban Flooding
## 62 21
## Urban/Small Stream Flood Urban/Sml Stream Fld
## 6 702
## Volcanic Ash Warm Weather
## 2 1
## Waterspout Waterspout-Tornado
## 30 1
## Waterspout Tornado Waterspout/ Tornado
## 1 1
## Waterspout/Tornado Wet Microburst
## 2 3
## Whirlwind Wild Fires
## 3 2
## Wild/Forest Fire Wild/Forest Fires
## 383 1
## Wildfire Wind
## 853 73
## Wind And Wave Wind Damage
## 1 8
## Wind Storm Winds
## 1 6
## Winter Storm Winter Storm High Winds
## 1479 1
## Winter Weather Winter Weather Mix
## 407 2
## Winter Weather/Mix Wintry Mix
## 139 4## Looking for events that may not have high frequency but which did cause big damage
## Create summary dataframe
EvTypeSummary <-
StormData %>%
group_by(EVTYPE) %>%
summarise(Freq = n(),
FATALITIES_TOTAL = sum(FATALITIES),
INJURIES_TOTAL = sum(INJURIES),
PROPDMG_TOTAL = sum(PROPDMG_CLEAN),
CROPDMG_TOTAL = sum(CROPDMG_CLEAN)
)Following the above (and partially informed by the stringdist exercise below) a number of replacements were carried out according to observed rules. e.g. TSTM shorthand for Thunderstorm replaced with the full name to standardise the values in this variable.
## Make a few known changes
# TSTM wind
StormData$EVTYPE[grepl("TSTM", StormData$EVTYPE, ignore.case = TRUE)] <- "Thunderstorm Wind"
# Landslide
StormData$EVTYPE[grepl("landslide", StormData$EVTYPE, ignore.case = TRUE)] <- "Debris Flow"
# Glaze
StormData$EVTYPE[grepl("glaze", StormData$EVTYPE, ignore.case = TRUE)] <- "Winter Storm"
# Thunderstorm wind - use "storm wind"
StormData$EVTYPE[grepl("storm wind", StormData$EVTYPE, ignore.case = TRUE)] <- "Thunderstorm Wind"
# Hurricane/typhoon
StormData$EVTYPE[grepl("hurricane", StormData$EVTYPE, ignore.case = TRUE)] <- "Hurricane (Typhoon)"
StormData$EVTYPE[grepl("typhoon", StormData$EVTYPE, ignore.case = TRUE)] <- "Hurricane (Typhoon)"
# Hail
StormData$EVTYPE[grepl("hail", StormData$EVTYPE, ignore.case = TRUE)] <- "Hail"
# Tropical storm
StormData$EVTYPE[grepl("tropical storm", StormData$EVTYPE, ignore.case = TRUE)] <- "Tropical Storm"
# Heat
StormData$EVTYPE[grepl("heat", StormData$EVTYPE, ignore.case = TRUE)] <- "Excessive Heat"
# Fire -> wildfire
StormData$EVTYPE[grepl("fire", StormData$EVTYPE, ignore.case = TRUE)] <- "Wildfire"
# Excessive snow -> heavy snow
StormData$EVTYPE[grepl("Excessive snow", StormData$EVTYPE, ignore.case = TRUE)] <- "Heavy snow"
# Light snow -> winter storm
StormData$EVTYPE[grepl("Light snow", StormData$EVTYPE, ignore.case = TRUE)] <- "Winter storm"
# Surf -> high surf
StormData$EVTYPE[grepl("surf", StormData$EVTYPE, ignore.case = TRUE)] <- "High surf"
## heavy seas -> Marine Strong Wind
StormData$EVTYPE[grepl("heavy sea", StormData$EVTYPE, ignore.case = TRUE)] <- "Marine Strong Wind"
## high water -> marine high wind.
StormData$EVTYPE[grepl("high water", StormData$EVTYPE, ignore.case = TRUE)] <- "Marine High Wind"
## Find unique values of EVTYPE
EVTYPEunique <- unique(StormData$EVTYPE)
## Check summary of EVTYPE frequencies
table(StormData$EVTYPE)##
## Agricultural Freeze Astronomical High Tide
## 3 8
## Astronomical Low Tide Avalanche
## 2 266
## Beach Erosion Black Ice
## 1 1
## Blizzard Blowing Dust
## 230 1
## Blowing Snow Breakup Flooding
## 2 1
## Coastal Erosion Coastal Flood
## 1 156
## Coastal Flooding Coastal Flooding/Erosion
## 36 4
## Coastal Storm Coastalstorm
## 4 1
## Cold Cold And Snow
## 28 1
## Cold And Wet Conditions Cold Temperature
## 1 2
## Cold Wave Cold Weather
## 1 4
## Cold/Wind Chill Dam Break
## 90 2
## Damaging Freeze Debris Flow
## 6 194
## Dense Fog Dense Smoke
## 64 1
## Downburst Drought
## 1 264
## Drowning Dry Microburst
## 1 77
## Dust Devil Dust Devil Waterspout
## 90 1
## Dust Storm Early Frost
## 98 1
## Erosion/Cstl Flood Excessive Heat
## 2 960
## Excessive Rainfall Excessive Wetness
## 1 1
## Extended Cold Extreme Cold
## 1 179
## Extreme Cold/Wind Chill Extreme Wind Chill
## 111 1
## Extreme Windchill Falling Snow/Ice
## 19 2
## Flash Flood Flash Flood - Heavy Rain
## 19851 2
## Flash Flood Winds Flash Flood/ Street
## 1 1
## Flash Flood/Flood Flash Flooding
## 3 100
## Flash Floods Flood
## 8 9712
## Flood & Heavy Rain Flood/Flash
## 2 1
## Flood/Flash Flood Flood/Flash/Flood
## 88 1
## Flood/Rain/Winds Flooding
## 5 23
## Flooding/Heavy Rain Floods
## 1 1
## Fog Freeze
## 104 16
## Freezing Drizzle Freezing Fog
## 7 7
## Freezing Rain Freezing Rain/Snow
## 18 3
## Freezing Spray Frost
## 1 2
## Frost/Freeze Funnel Cloud
## 117 9
## Gradient Wind Gustnado
## 6 2
## Gusty Wind Gusty Wind/Hvy Rain
## 13 1
## Gusty Wind/Rain Gusty Winds
## 1 30
## Hail Hard Freeze
## 23891 2
## Heavy Mix Heavy Rain
## 2 1080
## Heavy Rain And Flood Heavy Rain/Severe Weather
## 1 1
## Heavy Rains Heavy snow
## 4 25
## Heavy Snow Heavy Snow-Squalls
## 1134 15
## Heavy Snow And High Winds Heavy Snow And Strong Winds
## 1 1
## Heavy Snow Shower Heavy Snow Squalls
## 1 12
## Heavy Snow/High Winds & Flood Heavy Snow/Ice
## 1 1
## Heavy Swells High Seas
## 1 6
## High surf High Swells
## 221 2
## High Waves High Wind
## 1 5416
## High Wind (G40) High Wind Damage
## 2 2
## High Winds High Winds Heavy Rains
## 312 1
## Hurricane (Typhoon) Hvy Rain
## 228 1
## Hyperthermia/Exposure Hypothermia/Exposure
## 1 6
## Ice Ice Jam Flood (Minor
## 2 1
## Ice Jam Flooding Ice On Road
## 2 1
## Ice Roads Ice Storm
## 1 653
## Ice/Strong Winds Icy Roads
## 1 22
## Lake-Effect Snow Lake Effect Snow
## 194 4
## Lake Flood Lakeshore Flood
## 1 5
## Landslump Landspout
## 1 2
## Late Season Snow Light Freezing Rain
## 1 22
## Lightning Lightning And Heavy Rain
## 12026 1
## Lightning And Thunderstorm Win Lightning Injury
## 1 1
## Lightning Wauseon Lightning.
## 1 1
## Ligntning Marine Accident
## 1 1
## Marine High Wind Marine Mishap
## 21 2
## Marine Strong Wind Microburst
## 48 1
## Mixed Precip Mixed Precipitation
## 6 18
## Mud Slide Mud Slides Urban Flooding
## 2 1
## Mudslide Mudslides
## 5 1
## Non-Severe Wind Damage Other
## 1 34
## Rain Rain/Snow
## 3 2
## Rapidly Rising Water Record Cold
## 1 1
## Rip Current Rip Currents
## 389 241
## River And Stream Flood River Flood
## 2 95
## River Flooding Rock Slide
## 14 1
## Rogue Wave Rough Seas
## 1 3
## Rural Flood Seiche
## 1 9
## Severe Thunderstorm Severe Thunderstorms
## 5 2
## Sleet Snow
## 1 31
## Snow And Ice Snow Freezing Rain
## 3 10
## Snow Squall Snow Squalls
## 3 10
## Snow/Ice Snow/Sleet/Freezing Rain
## 1 2
## Storm Force Winds Storm Surge
## 1 169
## Storm Surge/Tide Strong Wind
## 47 3370
## Strong Winds Thundersnow
## 48 1
## Thunderstorm Thunderstorm Damage To
## 14 1
## Thunderstorm Wind Thunderstorm Wins
## 110907 1
## Thunderstorms Wind Thunderstorms Winds
## 1 4
## Thunderstormw Thundertorm Winds
## 1 3
## Tidal Flooding Tornado
## 4 13000
## Tornado F0 Tornado F1
## 5 4
## Tornado F2 Tornado F3
## 2 2
## Torrential Rainfall Tropical Depression
## 1 35
## Tropical Storm Tsunami
## 415 14
## Unseasonable Cold Unseasonably Cold
## 1 4
## Unseasonably Warm Unseasonably Warm And Dry
## 7 1
## Unseasonal Rain Urban And Small Stream Floodin
## 2 1
## Urban Flood Urban Flooding
## 62 21
## Urban/Small Stream Flood Urban/Sml Stream Fld
## 6 702
## Volcanic Ash Warm Weather
## 2 1
## Waterspout Waterspout-Tornado
## 30 1
## Waterspout Tornado Waterspout/ Tornado
## 1 1
## Waterspout/Tornado Wet Microburst
## 2 3
## Whirlwind Wildfire
## 3 1242
## Wind Wind And Wave
## 73 1
## Wind Damage Wind Storm
## 8 1
## Winds Winter storm
## 6 142
## Winter Storm Winter Storm High Winds
## 1501 1
## Winter Weather Winter Weather Mix
## 407 2
## Winter Weather/Mix Wintry Mix
## 139 4## Create summary dataframe
EvTypeSummary <-
StormData %>%
group_by(EVTYPE) %>%
summarise(Freq = n())This next chunk uses the stringdist_join function, as part of the stringdist package to determine the string distance between each original EVTYPE value and those in the official list. Those with distances less than a certain value are assigned a value from that official list. The Jaro Winkler distance method was used because the differences are due to typos rather than spoken words being mis-transcribed: https://cran.r-project.org/web/packages/stringdist/stringdist.pdf
This exercise assigned new values from the official list to 210332 records from the original 211775 (99%) although as noted earlier, these 99% will not all have been assigned correctly.
## Create new dataframe with new variables for EVTYPE, informed by stringdist matching.
StormDataDist <- stringdist_join(StormData, StormEventsDataFrame,
by = "EVTYPE",
mode = "left",
ignore_case = TRUE,
method = "jw",
max_dist = 0.2444444, # Informed by manually investigating first pass to find cut-off where matching appears inaccurate
distance_col = "dist") %>%
group_by(EVTYPE.x) %>%
slice_min(order_by = dist, n = 1)
## Investigating results of stringdist_join
print(paste0("Number of records in new StormDataDist dataframe: ", nrow(StormDataDist)))## [1] "Number of records in new StormDataDist dataframe: 210332"# Number of unique evtype values
length(unique(StormDataDist$EVTYPE.x))## [1] 137length(unique(StormDataDist$EVTYPE.y))## [1] 46# Number of unique events which could not find a perfect match
length(unique(StormDataDist$EVTYPE.x[StormDataDist$dist!=0]))## [1] 90# Frequencies of EVTYPE in original and new variable
table(StormDataDist$EVTYPE.x)##
## Astronomical High Tide Astronomical Low Tide
## 8 2
## Avalanche Blizzard
## 266 230
## Coastal Erosion Coastal Flood
## 1 156
## Coastal Flooding Coastal Flooding/Erosion
## 36 4
## Coastal Storm Coastalstorm
## 4 1
## Cold/Wind Chill Debris Flow
## 90 194
## Dense Fog Dense Smoke
## 64 1
## Drought Dust Devil
## 264 90
## Dust Devil Waterspout Dust Storm
## 1 98
## Excessive Heat Excessive Rainfall
## 960 1
## Excessive Wetness Extreme Cold
## 1 179
## Extreme Cold/Wind Chill Extreme Wind Chill
## 111 1
## Extreme Windchill Flash Flood
## 19 19851
## Flash Flood - Heavy Rain Flash Flood Winds
## 2 1
## Flash Flood/ Street Flash Flood/Flood
## 1 3
## Flash Flooding Flash Floods
## 100 8
## Flood Flood & Heavy Rain
## 9712 2
## Flood/Flash Flood/Flash Flood
## 1 88
## Flood/Flash/Flood Flood/Rain/Winds
## 1 5
## Flooding Floods
## 23 1
## Freezing Drizzle Freezing Fog
## 7 7
## Freezing Rain Freezing Rain/Snow
## 18 3
## Freezing Spray Frost
## 1 2
## Frost/Freeze Funnel Cloud
## 117 9
## Gustnado Hail
## 2 23891
## Heavy Mix Heavy Rain
## 2 1080
## Heavy Rain And Flood Heavy Rain/Severe Weather
## 1 1
## Heavy Rains Heavy snow
## 4 25
## Heavy Snow Heavy Snow-Squalls
## 1134 15
## Heavy Snow And High Winds Heavy Snow And Strong Winds
## 1 1
## Heavy Snow Shower Heavy Snow Squalls
## 1 12
## Heavy Snow/High Winds & Flood Heavy Snow/Ice
## 1 1
## Heavy Swells High Seas
## 1 6
## High surf High Wind
## 221 5416
## High Wind (G40) High Wind Damage
## 2 2
## High Winds High Winds Heavy Rains
## 312 1
## Hurricane (Typhoon) Hvy Rain
## 228 1
## Ice Ice Storm
## 2 653
## Ice/Strong Winds Lake-Effect Snow
## 1 194
## Lake Effect Snow Lake Flood
## 4 1
## Lakeshore Flood Lightning
## 5 12026
## Lightning And Heavy Rain Lightning And Thunderstorm Win
## 1 1
## Lightning Injury Lightning Wauseon
## 1 1
## Lightning. Ligntning
## 1 1
## Marine Accident Marine High Wind
## 1 21
## Marine Mishap Marine Strong Wind
## 2 48
## Rip Current Rip Currents
## 389 241
## River Flooding Rural Flood
## 14 1
## Seiche Sleet
## 9 1
## Storm Surge Storm Surge/Tide
## 169 47
## Strong Wind Strong Winds
## 3370 48
## Thundersnow Thunderstorm
## 1 14
## Thunderstorm Damage To Thunderstorm Wind
## 1 110907
## Thunderstorm Wins Thunderstorms Wind
## 1 1
## Thunderstorms Winds Thunderstormw
## 4 1
## Thundertorm Winds Tornado
## 3 13000
## Tornado F0 Tornado F1
## 5 4
## Tornado F2 Tornado F3
## 2 2
## Tropical Depression Tropical Storm
## 35 415
## Tsunami Urban Flood
## 14 62
## Urban Flooding Volcanic Ash
## 21 2
## Waterspout Waterspout-Tornado
## 30 1
## Waterspout Tornado Waterspout/ Tornado
## 1 1
## Waterspout/Tornado Whirlwind
## 2 3
## Wildfire Wind Storm
## 1242 1
## Winter storm Winter Storm
## 142 1501
## Winter Storm High Winds Winter Weather
## 1 407
## Winter Weather Mix Winter Weather/Mix
## 2 139
## Wintry Mix
## 4table(StormDataDist$EVTYPE.y)##
## Astronomical Low Tide Avalanche Blizzard
## 10 266 230
## Coastal Flood Cold/Wind Chill Debris Flow
## 198 90 194
## Dense Fog Dense Smoke Drought
## 64 1 264
## Dust Devil Dust Storm Excessive Heat
## 91 102 962
## Extreme Cold/Wind Chill Flash Flood Flood
## 310 20029 9868
## Freezing Fog Frost/Freeze Funnel Cloud
## 36 119 9
## Hail Heavy Rain Heavy Snow
## 23891 1089 1192
## High Surf High Wind Hurricane (Typhoon)
## 227 5736 228
## Ice Storm Lake-Effect Snow Lakeshore Flood
## 656 198 5
## Lightning Marine Hail Marine High Wind
## 12032 3 21
## Marine Strong Wind Rip Current Seiche
## 48 630 9
## Sleet Storm Surge/Tide Strong Wind
## 1 216 3419
## Thunderstorm Wind Tornado Tropical Depression
## 110933 13015 35
## Tropical Storm Tsunami Volcanic Ash
## 416 14 2
## Waterspout Wildfire Winter Storm
## 35 1242 1648
## Winter Weather
## 548# Find duplicates
length(StormDataDist$REFNUM[length(StormDataDist$REFNUM)>1]) == nrow(StormDataDist) # TRUE if no duplicates## [1] TRUEprint(StormDataDist[duplicated(StormDataDist$REFNUM), ]) # If there are any duplicates## # A tibble: 0 x 9
## # Groups: EVTYPE.x [0]
## # ... with 9 variables: EVTYPE.x <chr>, FATALITIES <dbl>, INJURIES <dbl>,
## # REFNUM <dbl>, BGN_DATE_CLEAN <date>, PROPDMG_CLEAN <dbl>,
## # CROPDMG_CLEAN <dbl>, EVTYPE.y <chr>, dist <dbl>The StormDataDist dataframe is then cleaned up further to remove the original EVTYPE variable and sum the fatality, injury and damage values per new EVTYPE.
The new dataframe is then used for analysis and results.
## Rename event type
StormDataDist <- rename(StormDataDist, EventType = EVTYPE.y)
## Ungroup
StormDataDist <- ungroup(StormDataDist)
## Remove unnecessary variables and summarise frequencies
StormDataCleanSummary <-
StormDataDist %>%
select(-EVTYPE.x, -REFNUM, -BGN_DATE_CLEAN, -dist) %>%
group_by(EventType) %>%
summarise(Freq = n(),
FATALITIES_TOTAL = sum(FATALITIES),
INJURIES_TOTAL = sum(INJURIES),
PROPDMG_TOTAL = sum(PROPDMG_CLEAN)/10^9,
CROPDMG_TOTAL = sum(CROPDMG_CLEAN)/10^9,
FATALITIES_MEAN = mean(FATALITIES),
INJURIES_MEAN = mean(INJURIES),
PROPDMG_MEAN = mean(PROPDMG_CLEAN)/10^6,
CROPDMG_MEAN = mean(CROPDMG_CLEAN)/10^6
)
## Remove unnecessary variables but no summary
StormDataClean <-
StormDataDist %>%
select(-EVTYPE.x, -REFNUM, -BGN_DATE_CLEAN, -dist)
## Print table (original sort, alphabetical)
print(StormDataClean)## # A tibble: 210,332 x 5
## FATALITIES INJURIES PROPDMG_CLEAN CROPDMG_CLEAN EventType
## <dbl> <dbl> <dbl> <dbl> <chr>
## 1 0 0 1000000 0 Astronomical Low Tide
## 2 0 0 5000000 0 Astronomical Low Tide
## 3 0 0 90000 0 Astronomical Low Tide
## 4 0 0 2500000 0 Astronomical Low Tide
## 5 0 0 20000 0 Astronomical Low Tide
## 6 0 0 800000 0 Astronomical Low Tide
## 7 0 0 5000 0 Astronomical Low Tide
## 8 0 0 10000 0 Astronomical Low Tide
## 9 0 0 120000 0 Astronomical Low Tide
## 10 0 0 200000 0 Astronomical Low Tide
## # ... with 210,322 more rowsResults
The following section presents the results for the two key questions.
The mean and sum (total) statistics were chosen to give an indication of impact per event and across the dataset of recorded events. These results are displayed below for fatalities, injuries, property damage and crop damage.
Most harmful weather events in the United States with respect to population health
## Fatalities
## Looking at mean
## Create initial ggplot object
PopHealthFatalityMean <-
ggplot(
slice_max(StormDataCleanSummary, order_by = FATALITIES_MEAN, n=10),
mapping = aes(reorder(EventType, FATALITIES_MEAN), FATALITIES_MEAN),
)
## Create plot
FatalityMeanPlot <-
PopHealthFatalityMean +
geom_bar(stat = "identity", fill = "indianred3") +
labs(y = "Mean Fatalities", x = "Event") +
geom_text(aes(label=round(FATALITIES_MEAN, 2)),
position=position_dodge(width=0.9),
hjust=-0.1
) +
coord_flip() +
ylim(0, 4)
## Looking at totals
## Create initial ggplot object
PopHealthFatalityTotal <-
ggplot(
slice_max(StormDataCleanSummary, order_by = FATALITIES_TOTAL, n=10),
mapping = aes(reorder(EventType, FATALITIES_TOTAL), FATALITIES_TOTAL)
)
## Create plot
FatalityTotalPlot <-
PopHealthFatalityTotal +
geom_bar(stat = "identity", fill = "indianred3") +
labs(y = "Total Fatalities", x = "Event") +
geom_text(aes(label=round(FATALITIES_TOTAL, 2)),
position=position_dodge(width=0.9)
) +
coord_flip() +
ylim(0, 4000)
## Injuries
## Looking at mean
## Create initial ggplot object
PopHealthInjuryMean <-
ggplot(
slice_max(StormDataCleanSummary, order_by = INJURIES_MEAN, n=10),
mapping = aes(reorder(EventType, INJURIES_MEAN), INJURIES_MEAN),
)
## Create plot
InjuryMeanPlot <-
PopHealthInjuryMean +
geom_bar(stat = "identity", fill = "orangered1") +
labs(y = "Mean Injuries", x = "Event") +
geom_text(aes(label=round(INJURIES_MEAN, 2)),
position=position_dodge(width=0.9)
) +
coord_flip()+
ylim(0, 10)
## Looking at totals
## Create initial ggplot object
PopHealthInjuryTotal <-
ggplot(
slice_max(StormDataCleanSummary, order_by = INJURIES_TOTAL, n=10),
mapping = aes(reorder(EventType, INJURIES_TOTAL), INJURIES_TOTAL)
)
## Create plot
InjuryTotalPlot <-
PopHealthInjuryTotal +
geom_bar(stat = "identity", fill = "orangered1") +
labs(y = "Total Injuries", x = "Event") +
geom_text(aes(label=round(INJURIES_TOTAL, 2)),
position=position_dodge(width=0.9)
) +
coord_flip()+
ylim(0, 25000)
## Arrange above plots into 2x2 grid
grid.arrange(FatalityMeanPlot, FatalityTotalPlot,InjuryMeanPlot, InjuryTotalPlot,
ncol=2,
nrow=2,
top = "Top 10 storm events"
)
Excessive heat appears to be the most harmful event in terms of fatalities per single event (mean) and across all recorded events (total). However individual events such as marine hail, tsunamis and sleet have the potential to cause as many fatalities per event.
In terms of injuries, excessive heat again causes the most injuries per event (mean), although tsunamis are not far behind. However, tornadoes come out on top in terms of total injuries. Further work is required with regards to tornadoes to determine if an outlier is contributing to this extremely high total number. That it is much lower in the mean chart suggests that there have just been more recorded Tornadoes.
Most harmful weather events in the United States with respect to Economic consequences
## Property damage
## Looking at mean
## Create initial ggplot object
EconPropMean <-
ggplot(
slice_max(StormDataCleanSummary, order_by = PROPDMG_MEAN, n=10),
mapping = aes(reorder(EventType, PROPDMG_MEAN), PROPDMG_MEAN),
)
## Create plot
EconPropMeanPlot <-
EconPropMean +
geom_bar(stat = "identity", fill = "cornflowerblue") +
labs(y = "Mean Property Damage ($M)", x = "Event") +
geom_text(aes(label=round(PROPDMG_MEAN, 2)),
position=position_dodge(width=0.9),
hjust=-0.1
) +
coord_flip() +
ylim(0, 500)
## Looking at totals
## Create initial ggplot object
EconPropTotal <-
ggplot(
slice_max(StormDataCleanSummary, order_by = PROPDMG_TOTAL, n=10),
mapping = aes(reorder(EventType, PROPDMG_TOTAL), PROPDMG_TOTAL)
)
## Create plot
EconPropTotalPlot <-
EconPropTotal +
geom_bar(stat = "identity", fill = "cornflowerblue") +
labs(y = "Total Property Damage ($B)", x = "Event") +
geom_text(aes(label=round(PROPDMG_TOTAL, 2)),
position=position_dodge(width=0.9)
) +
coord_flip() +
ylim(0, 200)
## Crop damage
## Looking at mean
## Create initial ggplot object
EconCropMean <-
ggplot(
slice_max(StormDataCleanSummary, order_by = CROPDMG_MEAN, n=10),
mapping = aes(reorder(EventType, CROPDMG_MEAN), CROPDMG_MEAN),
)
## Create plot
EconCropMeanPlot <-
EconCropMean +
geom_bar(stat = "identity", fill = "gold2") +
labs(y = "Mean Crop Damage ($M)", x = "Event") +
geom_text(aes(label=round(CROPDMG_MEAN, 2)),
position=position_dodge(width=0.9)
) +
coord_flip()+
ylim(0, 60)
## Looking at totals
## Create initial ggplot object
EconCropTotal <-
ggplot(
slice_max(StormDataCleanSummary, order_by = CROPDMG_TOTAL, n=10),
mapping = aes(reorder(EventType, CROPDMG_TOTAL), CROPDMG_TOTAL)
)
## Create plot
EconCropTotalPlot <-
EconCropTotal +
geom_bar(stat = "identity", fill = "gold2") +
labs(y = "Total Crop Damage ($B)", x = "Event") +
geom_text(aes(label=round(CROPDMG_TOTAL, 2)),
position=position_dodge(width=0.9)
) +
coord_flip()+
ylim(0, 20)
## Arrange above plots into 2x2 grid
grid.arrange(EconPropMeanPlot, EconPropTotalPlot,EconCropMeanPlot, EconCropTotalPlot,
ncol=2,
nrow=2,
top = "Top 10 storm events"
)
Hurricanes/typhoons are the most destructive in terms of property damage, per single event, but floods have caused more damage over the recorded events.
Droughts are by far the most destructive for crops, per single event and across all events.