title: “NOAA Storm Database Analysis”
author: “Katie Scoones”
date: “November 2026”
output: html_document
This project analyzes the National Oceanic and Atmospheric Administration (NOAA) Storm Events Database for 2024. The purpose of the analysis is to identify which types of storms have the greatest effects on public safety, occur most frequently across different states and months, and cause the most economic damage. The dataset contains 69,801 storm events across 50 different event types. The results show that tornadoes caused the most injuries, with 658 reported injuries. Excessive heat caused the most deaths, with 253 reported fatalities. Thunderstorm winds were the most frequently reported storm event overall. Winter weather was most common during the winter months, while thunderstorm winds were especially common during the warmer months. Hurricanes caused the greatest economic damage, totaling approximately $6.81 billion in property and crop damage. These findings can help government and emergency management officials better understand storm-related risks and prepare for future events.
The analysis used three CSV files from the NOAA Storm Events Database
for 2024: the storm details, the fatalities, and the locations files.
The files were downloaded and stored in a folder named Data
within the project folder.
The read.csv() function was used to import the three
files into R. The datasets were combined using the EVENT_ID
variable, which identifies individual storm events. A left join was used
to retain all records from the details dataset while adding matching
location and fatality information.
The combined dataset contained 86,556 rows and 70 columns. There were 69,801 unique event IDs in both the original details dataset and the combined dataset. Some storm events had multiple location or fatality records, which resulted in additional rows after joining.
For the main analysis, the original details dataset was used to calculate injuries, deaths, event frequencies, and economic damage. This was to avoid duplicating storm events because of the additional records created by the join functions.
The dataset contained 50 different event types. Checks to the main variables were also used in the analysis, including event type, state, month, injuries, and deaths, to see if there were any missing values. No missing values were found in these variables.
# Load required packages
library(dplyr)
library(readr)
library(ggplot2)
# Load the CSV files
details <- read.csv("Data/StormEvents_details-ftp_v1.0_d2024_c20260728.csv")
fatalities <- read.csv("Data/StormEvents_fatalities-ftp_v1.0_d2024_c20260728.csv")
locations <- read.csv("Data/StormEvents_locations-ftp_v1.0_d2024_c20260728.csv")
# Check the datasets
head(details)
## BEGIN_YEARMONTH BEGIN_DAY BEGIN_TIME END_YEARMONTH END_DAY END_TIME
## 1 202404 30 2033 202404 30 2033
## 2 202407 1 0 202407 5 900
## 3 202411 16 230 202411 18 1421
## 4 202405 22 1230 202405 22 1615
## 5 202405 21 1200 202405 21 1530
## 6 202405 24 1405 202405 24 1410
## EPISODE_ID EVENT_ID STATE STATE_FIPS YEAR MONTH_NAME EVENT_TYPE
## 1 189851 1174463 OKLAHOMA 40 2024 April Thunderstorm Wind
## 2 193486 1195301 LOUISIANA 22 2024 July Excessive Heat
## 3 197838 1223377 OREGON 41 2024 November Heavy Snow
## 4 191723 1184135 TEXAS 48 2024 May Heat
## 5 191723 1184133 TEXAS 48 2024 May Heat
## 6 191916 1182348 MISSISSIPPI 28 2024 May Thunderstorm Wind
## CZ_TYPE CZ_FIPS CZ_NAME WFO BEGIN_DATE_TIME
## 1 C 141 TILLMAN OUN 30-APR-24 20:33:00
## 2 Z 18 NATCHITOCHES SHV 01-JUL-24 00:00:00
## 3 Z 509 EAST SLOPES OF THE OREGON CASCADES PDT 16-NOV-24 02:30:00
## 4 Z 254 WILLACY BRO 22-MAY-24 12:30:00
## 5 Z 254 WILLACY BRO 21-MAY-24 12:00:00
## 6 C 115 PONTOTOC MEG 24-MAY-24 14:05:00
## CZ_TIMEZONE END_DATE_TIME INJURIES_DIRECT INJURIES_INDIRECT
## 1 CST-6 30-APR-24 20:33:00 0 0
## 2 CST-6 05-JUL-24 09:00:00 0 0
## 3 PST-8 18-NOV-24 14:21:00 0 0
## 4 CST-6 22-MAY-24 16:15:00 0 0
## 5 CST-6 21-MAY-24 15:30:00 0 0
## 6 CST-6 24-MAY-24 14:10:00 0 0
## DEATHS_DIRECT DEATHS_INDIRECT DAMAGE_PROPERTY DAMAGE_CROPS SOURCE MAGNITUDE
## 1 0 0 ASOS 55
## 2 0 0 0.00K 0.00K ASOS NA
## 3 0 0 0.00K 0.00K SNOTEL NA
## 4 0 0 0.00K 0.00K ASOS NA
## 5 0 0 0.00K 0.00K ASOS NA
## 6 0 0 1.00K 0.00K Public 52
## MAGNITUDE_TYPE FLOOD_CAUSE CATEGORY TOR_F_SCALE TOR_LENGTH TOR_WIDTH
## 1 MG NA NA NA
## 2 NA NA NA
## 3 NA NA NA
## 4 NA NA NA
## 5 NA NA NA
## 6 EG NA NA NA
## TOR_OTHER_WFO TOR_OTHER_CZ_STATE TOR_OTHER_CZ_FIPS TOR_OTHER_CZ_NAME
## 1 NA
## 2 NA
## 3 NA
## 4 NA
## 5 NA
## 6 NA
## BEGIN_RANGE BEGIN_AZIMUTH BEGIN_LOCATION END_RANGE END_AZIMUTH END_LOCATION
## 1 0.42 SSW FREDERICK ARPT 0.42 SSW FREDERICK ARPT
## 2 NA NA
## 3 NA NA
## 4 NA NA
## 5 NA NA
## 6 0.00 N ALGOMA 0.00 N ALGOMA
## BEGIN_LAT BEGIN_LON END_LAT END_LON
## 1 34.3444 -98.983 34.3444 -98.983
## 2 NA NA NA NA
## 3 NA NA NA NA
## 4 NA NA NA NA
## 5 NA NA NA NA
## 6 34.1800 -89.030 34.1800 -89.030
## EPISODE_NARRATIVE
## 1 A rather nebulous upper air pattern existed across the Southern Plains on the 30th, with 30-40 knots of quasi-zonal flow at 500 mb. However, plentiful solar insolation/surface heating in the vicinity of a dryline fostered widely scattered thunderstorm development by the late afternoon. Strong instability and sufficient organizing wind shear lead to a few of these storms displaying supercellular characteristics and producing reports of large hail initially, owing to steep lapse rates and weak low-level shear (among other factors already mentioned). By the mid-evening, as low-level shear and storm-scale interactions began to increase, a few tornadoes occurred across portions of south-central into southwest Oklahoma. The most significant of these was an EF-1 tornado near the community of Hollister (Tillman County).
## 2 An upper ridge of high pressure built in across the Southern Plains during the first several days of July, before eventually sliding off to the east into the Mid-South and Lower Mississippi Valley. Meanwhile, a stationary front lingered from Southeast Oklahoma into extreme Southwest Arkansas and portions of North Louisiana, which aided in compressional heating and low level moisture pooling, with temperatures climbing into the upper 90s to in excess of 100 degrees. The moisture pooling resulted in minimal mixing of dewpoints, with attendant heat indices ranging from 110-115 degrees each afternoon on July 1st through the 4th. However, slightly cooler and drier air were able to mix into the area on the 5th, resulting in slightly lower heat indices.
## 3 A series of cold fronts the weekend of Nov. 16th lowered snow levels across the region, which brought moderate to heavy snow accumulations across some mountain zones.
## 4 A strong upper-level subtropical ridge/heat dome centered over central and southern Mexico strengthened and shifted northward towards the Rio Grande Valley and Deep South Texas ranchlands the week prior to Memorial Day. The strong ridge aided lower level temperatures at around five thousand feet to soar to between 78 and 86F, which translated to surface temperatures rising into the upper 90s to mid 100s on most days between May 21-25. With dewpoint temperatures well into the 70s to around 80, apparent temperatures (feels like temperatures) ranged anywhere between 111-116F across parts of Deep South Texas and the Rio Grande Valley with a few locales reaching higher than this. The hottest apparent temperatures were located along and east of IH-69C from Hidalgo and Brooks County to Cameron, Willacy, and Kenedy County - but all of Deep South Texas observed at least dangerous heat levels by local definition (111F or higher for at least two hours) on most days.||Actual temperatures during the period were even hotter than the peak of the summer heat (late July through mid August) for most areas. The month would shatter all-time temperature records by 1 to more than 3 degrees.
## 5 A strong upper-level subtropical ridge/heat dome centered over central and southern Mexico strengthened and shifted northward towards the Rio Grande Valley and Deep South Texas ranchlands the week prior to Memorial Day. The strong ridge aided lower level temperatures at around five thousand feet to soar to between 78 and 86F, which translated to surface temperatures rising into the upper 90s to mid 100s on most days between May 21-25. With dewpoint temperatures well into the 70s to around 80, apparent temperatures (feels like temperatures) ranged anywhere between 111-116F across parts of Deep South Texas and the Rio Grande Valley with a few locales reaching higher than this. The hottest apparent temperatures were located along and east of IH-69C from Hidalgo and Brooks County to Cameron, Willacy, and Kenedy County - but all of Deep South Texas observed at least dangerous heat levels by local definition (111F or higher for at least two hours) on most days.||Actual temperatures during the period were even hotter than the peak of the summer heat (late July through mid August) for most areas. The month would shatter all-time temperature records by 1 to more than 3 degrees.
## 6 An active storm pattern as several mid-level shortwaves crossed the region. An MCV lifted east northeast out of the Ozarks across southern Arkansas. The MCV fired up morning convection across the Mid-South with isolated storm coverage in the afternoon. A weak capped environment was broken and steep mid-level lapse rates, roughly 30 knots of shear, and plentiful instability allowed storms to flourish.
## EVENT_NARRATIVE
## 1 Frederick Municipal Airport (KFDR) observation.
## 2
## 3 The Hog Pass SNOTEL reported an estimated 12 inches of new snow fell in a 24 hour period ending at 0800 PST on November 18th.
## 4 Harlingen Valley International Airport (KHRL) recorded max heat indices between 111-114F for nearly four hours from 130 to 515 PM CDT on May 22, 2024. By proxy, and data from northern Kenedy (1 hour of 111 to 112 heat index during this window), it can be inferred that inland Willacy County had similar heat index values for a similar period of time.
## 5 By proxy, between locations in northern Kenedy and in northern Cameron County, Willacy County is assumed to have observed heat index values between 111 and 114F between 1 and 430 PM CDT.
## 6 Reports of trees down in Algoma.
## DATA_SOURCE
## 1 CSV
## 2 CSV
## 3 CSV
## 4 CSV
## 5 CSV
## 6 CSV
head(fatalities)
## FAT_YEARMONTH FAT_DAY FAT_TIME FATALITY_ID EVENT_ID FATALITY_TYPE
## 1 202408 11 0 53724 1207220 D
## 2 202408 5 0 53725 1207865 I
## 3 202408 6 0 53726 1207865 I
## 4 202406 30 0 53169 1194107 D
## 5 202406 6 0 53180 1194997 D
## 6 202406 17 0 53181 1195001 D
## FATALITY_DATE FATALITY_AGE FATALITY_SEX FATALITY_LOCATION
## 1 08/11/2024 00:00:00 83 M In Water
## 2 08/05/2024 00:00:00 93 M Unknown
## 3 08/06/2024 00:00:00 63 F Unknown
## 4 06/30/2024 00:00:00 41 M In Water
## 5 06/06/2024 00:00:00 27 M Heavy Equipment/Construction
## 6 06/17/2024 00:00:00 16 M In Water
head(locations)
## YEARMONTH EPISODE_ID EVENT_ID LOCATION_INDEX RANGE AZIMUTH LOCATION
## 1 202407 194961 1205741 3 2.82 W TWENTY NINE PALMS
## 2 202407 194961 1205741 4 2.80 W TWENTY NINE PALMS
## 3 202407 194961 1205744 1 1.68 WNW TWENTY NINE PALMS
## 4 202407 194961 1205745 1 2.84 NNE LENWOOD
## 5 202407 194961 1205745 2 2.92 NNE LENWOOD
## 6 202407 194961 1205745 3 3.02 NNE LENWOOD
## LATITUDE LONGITUDE LAT2 LON2
## 1 34.1359 -116.0988 348154 1165928
## 2 34.1338 -116.0987 348028 1165922
## 3 34.1429 -116.0749 348574 1164494
## 4 34.9059 -117.0957 3454354 1175742
## 5 34.9059 -117.0929 3454354 1175574
## 6 34.9078 -117.0932 3454468 1175592
# Join datasets using EVENT_ID
joined_data <- details %>%
left_join(locations, by = "EVENT_ID") %>%
left_join(fatalities,
by = "EVENT_ID",
relationship = "many-to-many")
# Check the combined dataset
head(joined_data)
## BEGIN_YEARMONTH BEGIN_DAY BEGIN_TIME END_YEARMONTH END_DAY END_TIME
## 1 202404 30 2033 202404 30 2033
## 2 202407 1 0 202407 5 900
## 3 202411 16 230 202411 18 1421
## 4 202405 22 1230 202405 22 1615
## 5 202405 21 1200 202405 21 1530
## 6 202405 24 1405 202405 24 1410
## EPISODE_ID.x EVENT_ID STATE STATE_FIPS YEAR MONTH_NAME
## 1 189851 1174463 OKLAHOMA 40 2024 April
## 2 193486 1195301 LOUISIANA 22 2024 July
## 3 197838 1223377 OREGON 41 2024 November
## 4 191723 1184135 TEXAS 48 2024 May
## 5 191723 1184133 TEXAS 48 2024 May
## 6 191916 1182348 MISSISSIPPI 28 2024 May
## EVENT_TYPE CZ_TYPE CZ_FIPS CZ_NAME WFO
## 1 Thunderstorm Wind C 141 TILLMAN OUN
## 2 Excessive Heat Z 18 NATCHITOCHES SHV
## 3 Heavy Snow Z 509 EAST SLOPES OF THE OREGON CASCADES PDT
## 4 Heat Z 254 WILLACY BRO
## 5 Heat Z 254 WILLACY BRO
## 6 Thunderstorm Wind C 115 PONTOTOC MEG
## BEGIN_DATE_TIME CZ_TIMEZONE END_DATE_TIME INJURIES_DIRECT
## 1 30-APR-24 20:33:00 CST-6 30-APR-24 20:33:00 0
## 2 01-JUL-24 00:00:00 CST-6 05-JUL-24 09:00:00 0
## 3 16-NOV-24 02:30:00 PST-8 18-NOV-24 14:21:00 0
## 4 22-MAY-24 12:30:00 CST-6 22-MAY-24 16:15:00 0
## 5 21-MAY-24 12:00:00 CST-6 21-MAY-24 15:30:00 0
## 6 24-MAY-24 14:05:00 CST-6 24-MAY-24 14:10:00 0
## INJURIES_INDIRECT DEATHS_DIRECT DEATHS_INDIRECT DAMAGE_PROPERTY DAMAGE_CROPS
## 1 0 0 0
## 2 0 0 0 0.00K 0.00K
## 3 0 0 0 0.00K 0.00K
## 4 0 0 0 0.00K 0.00K
## 5 0 0 0 0.00K 0.00K
## 6 0 0 0 1.00K 0.00K
## SOURCE MAGNITUDE MAGNITUDE_TYPE FLOOD_CAUSE CATEGORY TOR_F_SCALE TOR_LENGTH
## 1 ASOS 55 MG NA NA
## 2 ASOS NA NA NA
## 3 SNOTEL NA NA NA
## 4 ASOS NA NA NA
## 5 ASOS NA NA NA
## 6 Public 52 EG NA NA
## TOR_WIDTH TOR_OTHER_WFO TOR_OTHER_CZ_STATE TOR_OTHER_CZ_FIPS
## 1 NA NA
## 2 NA NA
## 3 NA NA
## 4 NA NA
## 5 NA NA
## 6 NA NA
## TOR_OTHER_CZ_NAME BEGIN_RANGE BEGIN_AZIMUTH BEGIN_LOCATION END_RANGE
## 1 0.42 SSW FREDERICK ARPT 0.42
## 2 NA NA
## 3 NA NA
## 4 NA NA
## 5 NA NA
## 6 0.00 N ALGOMA 0.00
## END_AZIMUTH END_LOCATION BEGIN_LAT BEGIN_LON END_LAT END_LON
## 1 SSW FREDERICK ARPT 34.3444 -98.983 34.3444 -98.983
## 2 NA NA NA NA
## 3 NA NA NA NA
## 4 NA NA NA NA
## 5 NA NA NA NA
## 6 N ALGOMA 34.1800 -89.030 34.1800 -89.030
## EPISODE_NARRATIVE
## 1 A rather nebulous upper air pattern existed across the Southern Plains on the 30th, with 30-40 knots of quasi-zonal flow at 500 mb. However, plentiful solar insolation/surface heating in the vicinity of a dryline fostered widely scattered thunderstorm development by the late afternoon. Strong instability and sufficient organizing wind shear lead to a few of these storms displaying supercellular characteristics and producing reports of large hail initially, owing to steep lapse rates and weak low-level shear (among other factors already mentioned). By the mid-evening, as low-level shear and storm-scale interactions began to increase, a few tornadoes occurred across portions of south-central into southwest Oklahoma. The most significant of these was an EF-1 tornado near the community of Hollister (Tillman County).
## 2 An upper ridge of high pressure built in across the Southern Plains during the first several days of July, before eventually sliding off to the east into the Mid-South and Lower Mississippi Valley. Meanwhile, a stationary front lingered from Southeast Oklahoma into extreme Southwest Arkansas and portions of North Louisiana, which aided in compressional heating and low level moisture pooling, with temperatures climbing into the upper 90s to in excess of 100 degrees. The moisture pooling resulted in minimal mixing of dewpoints, with attendant heat indices ranging from 110-115 degrees each afternoon on July 1st through the 4th. However, slightly cooler and drier air were able to mix into the area on the 5th, resulting in slightly lower heat indices.
## 3 A series of cold fronts the weekend of Nov. 16th lowered snow levels across the region, which brought moderate to heavy snow accumulations across some mountain zones.
## 4 A strong upper-level subtropical ridge/heat dome centered over central and southern Mexico strengthened and shifted northward towards the Rio Grande Valley and Deep South Texas ranchlands the week prior to Memorial Day. The strong ridge aided lower level temperatures at around five thousand feet to soar to between 78 and 86F, which translated to surface temperatures rising into the upper 90s to mid 100s on most days between May 21-25. With dewpoint temperatures well into the 70s to around 80, apparent temperatures (feels like temperatures) ranged anywhere between 111-116F across parts of Deep South Texas and the Rio Grande Valley with a few locales reaching higher than this. The hottest apparent temperatures were located along and east of IH-69C from Hidalgo and Brooks County to Cameron, Willacy, and Kenedy County - but all of Deep South Texas observed at least dangerous heat levels by local definition (111F or higher for at least two hours) on most days.||Actual temperatures during the period were even hotter than the peak of the summer heat (late July through mid August) for most areas. The month would shatter all-time temperature records by 1 to more than 3 degrees.
## 5 A strong upper-level subtropical ridge/heat dome centered over central and southern Mexico strengthened and shifted northward towards the Rio Grande Valley and Deep South Texas ranchlands the week prior to Memorial Day. The strong ridge aided lower level temperatures at around five thousand feet to soar to between 78 and 86F, which translated to surface temperatures rising into the upper 90s to mid 100s on most days between May 21-25. With dewpoint temperatures well into the 70s to around 80, apparent temperatures (feels like temperatures) ranged anywhere between 111-116F across parts of Deep South Texas and the Rio Grande Valley with a few locales reaching higher than this. The hottest apparent temperatures were located along and east of IH-69C from Hidalgo and Brooks County to Cameron, Willacy, and Kenedy County - but all of Deep South Texas observed at least dangerous heat levels by local definition (111F or higher for at least two hours) on most days.||Actual temperatures during the period were even hotter than the peak of the summer heat (late July through mid August) for most areas. The month would shatter all-time temperature records by 1 to more than 3 degrees.
## 6 An active storm pattern as several mid-level shortwaves crossed the region. An MCV lifted east northeast out of the Ozarks across southern Arkansas. The MCV fired up morning convection across the Mid-South with isolated storm coverage in the afternoon. A weak capped environment was broken and steep mid-level lapse rates, roughly 30 knots of shear, and plentiful instability allowed storms to flourish.
## EVENT_NARRATIVE
## 1 Frederick Municipal Airport (KFDR) observation.
## 2
## 3 The Hog Pass SNOTEL reported an estimated 12 inches of new snow fell in a 24 hour period ending at 0800 PST on November 18th.
## 4 Harlingen Valley International Airport (KHRL) recorded max heat indices between 111-114F for nearly four hours from 130 to 515 PM CDT on May 22, 2024. By proxy, and data from northern Kenedy (1 hour of 111 to 112 heat index during this window), it can be inferred that inland Willacy County had similar heat index values for a similar period of time.
## 5 By proxy, between locations in northern Kenedy and in northern Cameron County, Willacy County is assumed to have observed heat index values between 111 and 114F between 1 and 430 PM CDT.
## 6 Reports of trees down in Algoma.
## DATA_SOURCE YEARMONTH EPISODE_ID.y LOCATION_INDEX RANGE AZIMUTH LOCATION
## 1 CSV NA NA NA NA <NA> <NA>
## 2 CSV NA NA NA NA <NA> <NA>
## 3 CSV NA NA NA NA <NA> <NA>
## 4 CSV NA NA NA NA <NA> <NA>
## 5 CSV NA NA NA NA <NA> <NA>
## 6 CSV NA NA NA NA <NA> <NA>
## LATITUDE LONGITUDE LAT2 LON2 FAT_YEARMONTH FAT_DAY FAT_TIME FATALITY_ID
## 1 NA NA NA NA NA NA NA NA
## 2 NA NA NA NA NA NA NA NA
## 3 NA NA NA NA NA NA NA NA
## 4 NA NA NA NA NA NA NA NA
## 5 NA NA NA NA NA NA NA NA
## 6 NA NA NA NA NA NA NA NA
## FATALITY_TYPE FATALITY_DATE FATALITY_AGE FATALITY_SEX FATALITY_LOCATION
## 1 <NA> <NA> NA <NA> <NA>
## 2 <NA> <NA> NA <NA> <NA>
## 3 <NA> <NA> NA <NA> <NA>
## 4 <NA> <NA> NA <NA> <NA>
## 5 <NA> <NA> NA <NA> <NA>
## 6 <NA> <NA> NA <NA> <NA>
# Check the dimensions of the combined dataset
dim(joined_data)
## [1] 86556 70
# Display the number of rows and columns
print(dim(joined_data))
## [1] 86556 70
# Count unique storm events in original details file
print(n_distinct(details$EVENT_ID))
## [1] 69801
# Count unique storm events in the combined dataset
print(n_distinct(joined_data$EVENT_ID))
## [1] 69801
# Display the column names
print(names(joined_data))
## [1] "BEGIN_YEARMONTH" "BEGIN_DAY" "BEGIN_TIME"
## [4] "END_YEARMONTH" "END_DAY" "END_TIME"
## [7] "EPISODE_ID.x" "EVENT_ID" "STATE"
## [10] "STATE_FIPS" "YEAR" "MONTH_NAME"
## [13] "EVENT_TYPE" "CZ_TYPE" "CZ_FIPS"
## [16] "CZ_NAME" "WFO" "BEGIN_DATE_TIME"
## [19] "CZ_TIMEZONE" "END_DATE_TIME" "INJURIES_DIRECT"
## [22] "INJURIES_INDIRECT" "DEATHS_DIRECT" "DEATHS_INDIRECT"
## [25] "DAMAGE_PROPERTY" "DAMAGE_CROPS" "SOURCE"
## [28] "MAGNITUDE" "MAGNITUDE_TYPE" "FLOOD_CAUSE"
## [31] "CATEGORY" "TOR_F_SCALE" "TOR_LENGTH"
## [34] "TOR_WIDTH" "TOR_OTHER_WFO" "TOR_OTHER_CZ_STATE"
## [37] "TOR_OTHER_CZ_FIPS" "TOR_OTHER_CZ_NAME" "BEGIN_RANGE"
## [40] "BEGIN_AZIMUTH" "BEGIN_LOCATION" "END_RANGE"
## [43] "END_AZIMUTH" "END_LOCATION" "BEGIN_LAT"
## [46] "BEGIN_LON" "END_LAT" "END_LON"
## [49] "EPISODE_NARRATIVE" "EVENT_NARRATIVE" "DATA_SOURCE"
## [52] "YEARMONTH" "EPISODE_ID.y" "LOCATION_INDEX"
## [55] "RANGE" "AZIMUTH" "LOCATION"
## [58] "LATITUDE" "LONGITUDE" "LAT2"
## [61] "LON2" "FAT_YEARMONTH" "FAT_DAY"
## [64] "FAT_TIME" "FATALITY_ID" "FATALITY_TYPE"
## [67] "FATALITY_DATE" "FATALITY_AGE" "FATALITY_SEX"
## [70] "FATALITY_LOCATION"
# Count the number of different event types in the combined dataset
length(unique(details$EVENT_TYPE))
## [1] 50
# Display the 15 most common event types of the combined dataset
sort(table(details$EVENT_TYPE), decreasing = TRUE)[1:15]
##
## Thunderstorm Wind Hail Flash Flood
## 19940 8941 4661
## High Wind Winter Weather Drought
## 3994 3557 3473
## Heat Excessive Heat Flood
## 3171 2660 2472
## Heavy Snow Marine Thunderstorm Wind Tornado
## 2210 2192 2137
## Winter Storm Strong Wind Heavy Rain
## 2112 1311 918
# Check for missing values in important columns
colSums(is.na(details[, c("EVENT_TYPE", "STATE", "MONTH_NAME",
"INJURIES_DIRECT", "INJURIES_INDIRECT",
"DEATHS_DIRECT", "DEATHS_INDIRECT")]))
## EVENT_TYPE STATE MONTH_NAME INJURIES_DIRECT
## 0 0 0 0
## INJURIES_INDIRECT DEATHS_DIRECT DEATHS_INDIRECT
## 0 0 0
To examine the effects of storms on population health, total injuries and deaths were calculated for each event type, including both direct and indirect injuries and fatalities.
In 2024, tornadoes caused the greatest number of injuries with 658 reported. Heat and thunderstorm wind followed, with 173 and 161 injuries.
With 253 fatalities, excessive heat caused the most deaths followed by heat with 219 deaths. Another significant cause of fatalities were flash floods with 117 deaths.
The results suggest that different types of storms create different risks. Tornadoes were particularly harmful in terms of injuries, while excessive heat and heat-related events were leading the causes of fatalities.
# Calculate total injuries and deaths for each event type
health_impact <- details %>%
group_by(EVENT_TYPE) %>%
summarize(
Total_Injuries = sum(INJURIES_DIRECT + INJURIES_INDIRECT),
Total_Deaths = sum(DEATHS_DIRECT + DEATHS_INDIRECT)
)
# Display the top 10 event types by injuries
health_impact %>%
arrange(desc(Total_Injuries)) %>%
slice_head(n = 10)
## # A tibble: 10 × 3
## EVENT_TYPE Total_Injuries Total_Deaths
## <chr> <int> <int>
## 1 Tornado 658 53
## 2 Heat 173 219
## 3 Thunderstorm Wind 161 53
## 4 Excessive Heat 100 253
## 5 Winter Weather 87 25
## 6 Hurricane (Typhoon) 85 17
## 7 Dust Storm 74 1
## 8 Dense Fog 65 18
## 9 Ice Storm 58 7
## 10 Wildfire 57 10
# Display the top 10 event types by deaths
health_impact %>%
arrange(desc(Total_Deaths)) %>%
slice_head(n = 10)
## # A tibble: 10 × 3
## EVENT_TYPE Total_Injuries Total_Deaths
## <chr> <int> <int>
## 1 Excessive Heat 100 253
## 2 Heat 173 219
## 3 Flash Flood 19 117
## 4 Tropical Storm 12 78
## 5 Rip Current 42 65
## 6 Thunderstorm Wind 161 53
## 7 Tornado 658 53
## 8 Cold/Wind Chill 0 39
## 9 Flood 0 35
## 10 Winter Weather 87 25
injury_plot <- health_impact %>%
arrange(desc(Total_Injuries)) %>%
slice_head(n = 10) %>%
ggplot(aes(x = reorder(EVENT_TYPE, Total_Injuries),
y = Total_Injuries)) +
geom_col(fill = "steelblue") +
geom_text(aes(label = Total_Injuries), hjust = -0.3) +
coord_flip(clip = "off") +
scale_y_continuous(expand = expansion(mult = c(0, 0.12))) +
labs(
title = "Top 10 Storm Events Causing Injuries",
x = "Storm Event Type",
y = "Total Injuries",
caption = "Source: NOAA Storm Events Database, 2024"
) +
theme_minimal()
print(injury_plot)
Figure 1. The ten storm event types with the highest total number of reported injuries in 2024.
death_plot <- health_impact %>%
arrange(desc(Total_Deaths)) %>%
slice_head(n = 10) %>%
ggplot(aes(x = reorder(EVENT_TYPE, Total_Deaths),
y = Total_Deaths)) +
geom_col(fill = "firebrick") +
geom_text(aes(label = Total_Deaths), hjust = -0.3) +
coord_flip(clip = "off") +
scale_y_continuous(expand = expansion(mult = c(0, 0.12))) +
labs(
title = "Top 10 Storm Events Causing Deaths",
x = "Storm Event Type",
y = "Total Deaths",
caption = "Source: NOAA Storm Events Database, 2024"
) +
theme_minimal()
print(death_plot)
Figure 2. The ten storm event types with the highest total number of reported deaths in 2024.
To examine storm events geographically, the number of reported events was calculated for each combination of state and event type. The most frequently reported event type was then identified for each state.
Thunderstorm wind was a leader in total events and exhibited the most in New York, Kansas, Georgia, Texas, and Pennsylvania. Hail was another leader of total events, and Texas, Kansas, Oklahoma, Nebraska, and Missouri exhibited the most. And for leading total events in Winter Weather, the most were reported in New York, Colorado, West Virginia, South Dakota, and Wisconsin.
These results demonstrate that storm patterns vary geographically. Some states experience more frequent thunderstorms, while others are more affected by winter weather, hail, or other regional hazards.
# Count storm events by state and event type
state_events <- details %>%
group_by(STATE, EVENT_TYPE) %>%
summarize(
Total_Events = n(),
.groups = "drop"
)
# Find the most common event type in each state
state_events %>%
group_by(STATE) %>%
slice_max(Total_Events, n = 1) %>%
arrange(STATE)
## # A tibble: 70 × 3
## # Groups: STATE [68]
## STATE EVENT_TYPE Total_Events
## <chr> <chr> <int>
## 1 ALABAMA Thunderstorm Wind 569
## 2 ALASKA Blizzard 85
## 3 AMERICAN SAMOA Flash Flood 6
## 4 AMERICAN SAMOA Heavy Rain 6
## 5 AMERICAN SAMOA Thunderstorm Wind 6
## 6 ARIZONA Thunderstorm Wind 174
## 7 ARKANSAS Thunderstorm Wind 329
## 8 ATLANTIC NORTH Marine Thunderstorm Wind 652
## 9 ATLANTIC SOUTH Marine Thunderstorm Wind 486
## 10 CALIFORNIA High Wind 583
## # ℹ 60 more rows
top_state_events <- state_events %>%
group_by(STATE) %>%
slice_max(Total_Events, n = 1) %>%
arrange(STATE)
# Find the top 20 states by total number of events
top_20_states <- state_events %>%
group_by(STATE) %>%
summarize(Total_State_Events = sum(Total_Events), .groups = "drop") %>%
arrange(desc(Total_State_Events)) %>%
slice_head(n = 20)
# Find the most common event type for each of the top 20 states
top_state_events <- state_events %>%
filter(STATE %in% top_20_states$STATE) %>%
group_by(STATE) %>%
slice_max(Total_Events, n = 1) %>%
arrange(desc(Total_Events))
# View the results
print(top_state_events, n = 20)
## # A tibble: 20 × 3
## # Groups: STATE [20]
## STATE EVENT_TYPE Total_Events
## <chr> <chr> <int>
## 1 TEXAS Hail 1588
## 2 KANSAS Thunderstorm Wind 1043
## 3 NEW YORK Thunderstorm Wind 1021
## 4 GEORGIA Thunderstorm Wind 1002
## 5 PENNSYLVANIA Thunderstorm Wind 971
## 6 ILLINOIS Thunderstorm Wind 959
## 7 OKLAHOMA Heat 912
## 8 KENTUCKY Thunderstorm Wind 818
## 9 VIRGINIA Thunderstorm Wind 732
## 10 MISSOURI Thunderstorm Wind 723
## 11 NORTH CAROLINA Thunderstorm Wind 696
## 12 IOWA Thunderstorm Wind 659
## 13 OHIO Thunderstorm Wind 639
## 14 NEBRASKA Thunderstorm Wind 599
## 15 CALIFORNIA High Wind 583
## 16 TENNESSEE Thunderstorm Wind 504
## 17 SOUTH DAKOTA Thunderstorm Wind 501
## 18 WEST VIRGINIA Thunderstorm Wind 481
## 19 COLORADO Heavy Snow 439
## 20 MINNESOTA Thunderstorm Wind 363
# Find the top 5 states for each event type
top_5_states_by_event <- state_events %>%
group_by(EVENT_TYPE) %>%
slice_max(Total_Events, n = 5) %>%
arrange(EVENT_TYPE, desc(Total_Events))
# Print the results
print(top_5_states_by_event, n = Inf)
## # A tibble: 222 × 3
## # Groups: EVENT_TYPE [50]
## STATE EVENT_TYPE Total_Events
## <chr> <chr> <int>
## 1 LOUISIANA Astronomical Low Tide 8
## 2 DELAWARE Astronomical Low Tide 2
## 3 TEXAS Astronomical Low Tide 2
## 4 ALABAMA Astronomical Low Tide 1
## 5 CONNECTICUT Astronomical Low Tide 1
## 6 NEW JERSEY Astronomical Low Tide 1
## 7 NEW YORK Astronomical Low Tide 1
## 8 TENNESSEE Astronomical Low Tide 1
## 9 UTAH Avalanche 10
## 10 IDAHO Avalanche 6
## 11 COLORADO Avalanche 4
## 12 WYOMING Avalanche 2
## 13 ALASKA Avalanche 1
## 14 NEVADA Avalanche 1
## 15 VIRGINIA Avalanche 1
## 16 WASHINGTON Avalanche 1
## 17 ALASKA Blizzard 85
## 18 IOWA Blizzard 64
## 19 NEBRASKA Blizzard 38
## 20 KANSAS Blizzard 29
## 21 NORTH DAKOTA Blizzard 25
## 22 NEW JERSEY Coastal Flood 77
## 23 MARYLAND Coastal Flood 51
## 24 VIRGINIA Coastal Flood 30
## 25 NEW YORK Coastal Flood 24
## 26 MAINE Coastal Flood 22
## 27 TEXAS Cold/Wind Chill 122
## 28 ILLINOIS Cold/Wind Chill 78
## 29 OKLAHOMA Cold/Wind Chill 78
## 30 MISSOURI Cold/Wind Chill 46
## 31 WISCONSIN Cold/Wind Chill 46
## 32 PUERTO RICO Debris Flow 78
## 33 CALIFORNIA Debris Flow 37
## 34 COLORADO Debris Flow 33
## 35 NEW MEXICO Debris Flow 6
## 36 IDAHO Debris Flow 4
## 37 CALIFORNIA Dense Fog 226
## 38 KENTUCKY Dense Fog 135
## 39 ILLINOIS Dense Fog 132
## 40 MISSOURI Dense Fog 96
## 41 NEW JERSEY Dense Fog 68
## 42 TEXAS Drought 449
## 43 NEW MEXICO Drought 375
## 44 IOWA Drought 299
## 45 TENNESSEE Drought 244
## 46 WEST VIRGINIA Drought 210
## 47 NEVADA Dust Devil 3
## 48 ARIZONA Dust Devil 1
## 49 CALIFORNIA Dust Devil 1
## 50 GUAM Dust Devil 1
## 51 OREGON Dust Devil 1
## 52 ARIZONA Dust Storm 28
## 53 NEW MEXICO Dust Storm 15
## 54 KANSAS Dust Storm 13
## 55 ILLINOIS Dust Storm 8
## 56 CALIFORNIA Dust Storm 7
## 57 OKLAHOMA Excessive Heat 567
## 58 TEXAS Excessive Heat 336
## 59 CALIFORNIA Excessive Heat 186
## 60 MISSISSIPPI Excessive Heat 163
## 61 PUERTO RICO Excessive Heat 158
## 62 NEBRASKA Extreme Cold/Wind Chill 101
## 63 IOWA Extreme Cold/Wind Chill 81
## 64 SOUTH DAKOTA Extreme Cold/Wind Chill 78
## 65 INDIANA Extreme Cold/Wind Chill 58
## 66 WYOMING Extreme Cold/Wind Chill 53
## 67 TEXAS Flash Flood 539
## 68 NORTH CAROLINA Flash Flood 367
## 69 VIRGINIA Flash Flood 281
## 70 LOUISIANA Flash Flood 244
## 71 GEORGIA Flash Flood 231
## 72 CALIFORNIA Flood 414
## 73 PUERTO RICO Flood 207
## 74 IOWA Flood 185
## 75 VIRGINIA Flood 169
## 76 MISSOURI Flood 130
## 77 ARKANSAS Freezing Fog 1
## 78 CALIFORNIA Freezing Fog 1
## 79 IDAHO Freezing Fog 1
## 80 MINNESOTA Freezing Fog 1
## 81 WASHINGTON Freezing Fog 1
## 82 CALIFORNIA Frost/Freeze 70
## 83 OREGON Frost/Freeze 61
## 84 TEXAS Frost/Freeze 47
## 85 KENTUCKY Frost/Freeze 46
## 86 ILLINOIS Frost/Freeze 43
## 87 IOWA Funnel Cloud 52
## 88 FLORIDA Funnel Cloud 34
## 89 NEBRASKA Funnel Cloud 26
## 90 NORTH DAKOTA Funnel Cloud 20
## 91 TEXAS Funnel Cloud 17
## 92 TEXAS Hail 1588
## 93 KANSAS Hail 767
## 94 OKLAHOMA Hail 765
## 95 NEBRASKA Hail 510
## 96 MISSOURI Hail 496
## 97 OKLAHOMA Heat 912
## 98 TEXAS Heat 828
## 99 MISSISSIPPI Heat 195
## 100 TENNESSEE Heat 146
## 101 NEW YORK Heat 132
## 102 VIRGINIA Heavy Rain 150
## 103 MISSOURI Heavy Rain 106
## 104 CALIFORNIA Heavy Rain 104
## 105 IOWA Heavy Rain 104
## 106 FLORIDA Heavy Rain 94
## 107 COLORADO Heavy Snow 439
## 108 CALIFORNIA Heavy Snow 275
## 109 NEVADA Heavy Snow 103
## 110 NEW MEXICO Heavy Snow 95
## 111 WASHINGTON Heavy Snow 90
## 112 HAWAII High Surf 77
## 113 CALIFORNIA High Surf 13
## 114 OREGON High Surf 11
## 115 PUERTO RICO High Surf 4
## 116 AMERICAN SAMOA High Surf 3
## 117 CALIFORNIA High Wind 583
## 118 MONTANA High Wind 435
## 119 TEXAS High Wind 318
## 120 WYOMING High Wind 271
## 121 NEVADA High Wind 245
## 122 FLORIDA Hurricane (Typhoon) 40
## 123 GEORGIA Hurricane (Typhoon) 29
## 124 LOUISIANA Hurricane (Typhoon) 7
## 125 IOWA Ice Storm 30
## 126 OREGON Ice Storm 20
## 127 WASHINGTON Ice Storm 20
## 128 ILLINOIS Ice Storm 11
## 129 TEXAS Ice Storm 11
## 130 NEW YORK Lake-Effect Snow 57
## 131 PENNSYLVANIA Lake-Effect Snow 32
## 132 OHIO Lake-Effect Snow 20
## 133 INDIANA Lake-Effect Snow 9
## 134 MICHIGAN Lake-Effect Snow 6
## 135 NEW YORK Lakeshore Flood 1
## 136 FLORIDA Lightning 23
## 137 TEXAS Lightning 18
## 138 TENNESSEE Lightning 14
## 139 GEORGIA Lightning 9
## 140 MISSISSIPPI Lightning 9
## 141 SOUTH CAROLINA Lightning 9
## 142 LAKE ERIE Marine Dense Fog 2
## 143 ATLANTIC SOUTH Marine Hail 9
## 144 LAKE ERIE Marine Hail 4
## 145 LAKE MICHIGAN Marine Hail 4
## 146 LAKE SUPERIOR Marine Hail 4
## 147 ATLANTIC NORTH Marine Hail 1
## 148 LAKE SUPERIOR Marine High Wind 32
## 149 ATLANTIC NORTH Marine High Wind 16
## 150 LAKE MICHIGAN Marine High Wind 14
## 151 ATLANTIC SOUTH Marine High Wind 1
## 152 LAKE ERIE Marine High Wind 1
## 153 GULF OF MEXICO Marine Hurricane/Typhoon 10
## 154 ATLANTIC SOUTH Marine Hurricane/Typhoon 1
## 155 GUAM WATERS Marine Strong Wind 1
## 156 GULF OF MEXICO Marine Strong Wind 1
## 157 LAKE ERIE Marine Strong Wind 1
## 158 GULF OF MEXICO Marine Thunderstorm Wind 728
## 159 ATLANTIC NORTH Marine Thunderstorm Wind 652
## 160 ATLANTIC SOUTH Marine Thunderstorm Wind 486
## 161 LAKE MICHIGAN Marine Thunderstorm Wind 117
## 162 LAKE ERIE Marine Thunderstorm Wind 92
## 163 ATLANTIC SOUTH Marine Tropical Depression 1
## 164 GULF OF MEXICO Marine Tropical Storm 45
## 165 ATLANTIC SOUTH Marine Tropical Storm 29
## 166 FLORIDA Rip Current 18
## 167 PUERTO RICO Rip Current 16
## 168 GUAM Rip Current 8
## 169 SOUTH CAROLINA Rip Current 7
## 170 NORTH CAROLINA Rip Current 6
## 171 ILLINOIS Seiche 2
## 172 INDIANA Seiche 1
## 173 ALABAMA Sleet 6
## 174 OREGON Sneakerwave 12
## 175 CALIFORNIA Sneakerwave 3
## 176 LOUISIANA Storm Surge/Tide 32
## 177 FLORIDA Storm Surge/Tide 28
## 178 MISSISSIPPI Storm Surge/Tide 6
## 179 TEXAS Storm Surge/Tide 5
## 180 WEST VIRGINIA Strong Wind 187
## 181 SOUTH CAROLINA Strong Wind 112
## 182 SOUTH DAKOTA Strong Wind 105
## 183 KENTUCKY Strong Wind 96
## 184 NEW YORK Strong Wind 87
## 185 KANSAS Thunderstorm Wind 1043
## 186 NEW YORK Thunderstorm Wind 1021
## 187 GEORGIA Thunderstorm Wind 1002
## 188 TEXAS Thunderstorm Wind 988
## 189 PENNSYLVANIA Thunderstorm Wind 971
## 190 OKLAHOMA Tornado 178
## 191 ILLINOIS Tornado 174
## 192 IOWA Tornado 155
## 193 TEXAS Tornado 149
## 194 FLORIDA Tornado 132
## 195 FLORIDA Tropical Depression 29
## 196 ALABAMA Tropical Depression 4
## 197 GUAM Tropical Depression 3
## 198 FLORIDA Tropical Storm 187
## 199 GEORGIA Tropical Storm 169
## 200 SOUTH CAROLINA Tropical Storm 59
## 201 NORTH CAROLINA Tropical Storm 35
## 202 LOUISIANA Tropical Storm 32
## 203 GULF OF MEXICO Waterspout 93
## 204 ATLANTIC SOUTH Waterspout 55
## 205 LAKE MICHIGAN Waterspout 19
## 206 LAKE ERIE Waterspout 11
## 207 E PACIFIC Waterspout 5
## 208 CALIFORNIA Wildfire 56
## 209 OKLAHOMA Wildfire 41
## 210 TEXAS Wildfire 37
## 211 OREGON Wildfire 28
## 212 COLORADO Wildfire 27
## 213 IOWA Winter Storm 175
## 214 MICHIGAN Winter Storm 156
## 215 WISCONSIN Winter Storm 144
## 216 NEBRASKA Winter Storm 138
## 217 WYOMING Winter Storm 103
## 218 NEW YORK Winter Weather 283
## 219 COLORADO Winter Weather 273
## 220 WEST VIRGINIA Winter Weather 194
## 221 SOUTH DAKOTA Winter Weather 185
## 222 WISCONSIN Winter Weather 176
# Find the top 5 most common event types overall
top_5_events <- state_events %>%
group_by(EVENT_TYPE) %>%
summarize(Total_Events = sum(Total_Events), .groups = "drop") %>%
arrange(desc(Total_Events)) %>%
slice_head(n = 5)
# Find the top 5 states for each of the top 5 event types
top_5_states_top_events <- state_events %>%
filter(EVENT_TYPE %in% top_5_events$EVENT_TYPE) %>%
group_by(EVENT_TYPE) %>%
slice_max(Total_Events, n = 5) %>%
arrange(EVENT_TYPE, desc(Total_Events))
# Print the results
print(top_5_states_top_events, n = Inf)
## # A tibble: 25 × 3
## # Groups: EVENT_TYPE [5]
## STATE EVENT_TYPE Total_Events
## <chr> <chr> <int>
## 1 TEXAS Flash Flood 539
## 2 NORTH CAROLINA Flash Flood 367
## 3 VIRGINIA Flash Flood 281
## 4 LOUISIANA Flash Flood 244
## 5 GEORGIA Flash Flood 231
## 6 TEXAS Hail 1588
## 7 KANSAS Hail 767
## 8 OKLAHOMA Hail 765
## 9 NEBRASKA Hail 510
## 10 MISSOURI Hail 496
## 11 CALIFORNIA High Wind 583
## 12 MONTANA High Wind 435
## 13 TEXAS High Wind 318
## 14 WYOMING High Wind 271
## 15 NEVADA High Wind 245
## 16 KANSAS Thunderstorm Wind 1043
## 17 NEW YORK Thunderstorm Wind 1021
## 18 GEORGIA Thunderstorm Wind 1002
## 19 TEXAS Thunderstorm Wind 988
## 20 PENNSYLVANIA Thunderstorm Wind 971
## 21 NEW YORK Winter Weather 283
## 22 COLORADO Winter Weather 273
## 23 WEST VIRGINIA Winter Weather 194
## 24 SOUTH DAKOTA Winter Weather 185
## 25 WISCONSIN Winter Weather 176
To examine seasonal patterns, the number of storm events was calculated for each month and event type. The most frequently reported event type for each month was also identified.
Winter weather was the most common event type in January, February, and December. Thunderstorm wind was the leading event type from April through September. Hail was the most common event type in March, while drought was most frequently reported in October and November.
The results show that storm activity changes throughout the year. Unsurprisingly, winter weather is more common during colder months, while thunderstorms and hail are more frequent during the warmer months. These patterns can help communities prepare for the types of severe weather they are more likely to experience during different seasons.
# Count storm events by month and event type
monthly_events <- details %>%
group_by(MONTH_NAME, EVENT_TYPE) %>%
summarize(Total_Events = n(), .groups = "drop")
# Find the most common event type in each month
monthly_events %>%
group_by(MONTH_NAME) %>%
slice_max(Total_Events, n = 1) %>%
arrange(match(MONTH_NAME, month.name))
## # A tibble: 12 × 3
## # Groups: MONTH_NAME [12]
## MONTH_NAME EVENT_TYPE Total_Events
## <chr> <chr> <int>
## 1 January Winter Weather 1584
## 2 February Winter Weather 625
## 3 March Hail 847
## 4 April Thunderstorm Wind 1441
## 5 May Thunderstorm Wind 4638
## 6 June Thunderstorm Wind 3895
## 7 July Thunderstorm Wind 4162
## 8 August Thunderstorm Wind 3126
## 9 September Thunderstorm Wind 666
## 10 October Drought 522
## 11 November Drought 648
## 12 December Winter Weather 584
# Identify the five most common event types overall
top_monthly_types <- details %>%
count(EVENT_TYPE, sort = TRUE) %>%
slice_head(n = 5) %>%
pull(EVENT_TYPE)
# Count monthly events for those five event types
monthly_plot_data <- details %>%
filter(EVENT_TYPE %in% top_monthly_types) %>%
group_by(MONTH_NAME, EVENT_TYPE) %>%
summarize(Total_Events = n(), .groups = "drop") %>%
mutate(Month_Number = match(MONTH_NAME, month.name))
# Create a line chart
monthly_plot <- ggplot(monthly_plot_data,
aes(x = Month_Number,
y = Total_Events,
color = EVENT_TYPE,
group = EVENT_TYPE)) +
geom_line(linewidth = 1) +
geom_point(size = 2) +
scale_x_continuous(
breaks = 1:12,
labels = month.abb
) +
labs(
title = "Monthly Patterns in the Five Most Common Storm Events",
x = "Month",
y = "Number of Events",
color = "Event Type",
caption = "Source: NOAA Storm Events Database, 2024"
) +
theme_minimal()
print(monthly_plot)
Figure 3. Monthly frequency of the five most commonly reported storm event types in 2024.
Lastly, we examined the economic effects of different events, both property and crop damage were included in this analysis.
The original damage variables were recorded as text, with suffixes representing thousands (K), millions (M), and billions (B) of dollars. These values were converted into numerical dollar amounts so that total economic damage could be calculated and properly compared.
In 2024, the results show that hurricanes caused the greatest economic damage, with approximately $6.81 billion in combined property and crop damage. Flash floods followed with about $3.24 billion, while tropical storms caused around $2.16 billion.
These findings demonstrate that certain storm events can have higher financial consequences. Hurricanes, floods, and tropical storms were especially costly, highlighting the importance of preparation and disaster planning to help reduce property and crop losses.
# Convert property and crop damage into dollar amounts
convert_damage <- function(x) {
x[x == ""] <- "0"
multiplier <- ifelse(grepl("K$", x), 1000,
ifelse(grepl("M$", x), 1000000,
ifelse(grepl("B$", x), 1000000000, 1)))
as.numeric(gsub("[KMB]", "", x)) * multiplier
}
# Create numeric damage columns
details$Property_Damage <- convert_damage(details$DAMAGE_PROPERTY)
details$Crop_Damage <- convert_damage(details$DAMAGE_CROPS)
# Calculate total damage by event type
economic_impact <- details %>%
group_by(EVENT_TYPE) %>%
summarize(
Total_Property_Damage = sum(Property_Damage),
Total_Crop_Damage = sum(Crop_Damage),
Total_Economic_Damage = sum(Property_Damage + Crop_Damage)
)
# Display the top 10 event types by total economic damage
economic_impact %>%
arrange(desc(Total_Economic_Damage)) %>%
slice_head(n = 10)
## # A tibble: 10 × 4
## EVENT_TYPE Total_Property_Damage Total_Crop_Damage Total_Economic_Damage
## <chr> <dbl> <dbl> <dbl>
## 1 Hurricane (Typ… 5302560000 1510000000 6812560000
## 2 Flash Flood 3143421200 101576500 3244997700
## 3 Tropical Storm 1902007000 253006000 2155013000
## 4 Tornado 1721030380 8531200 1729561580
## 5 Wildfire 1102899870 64817000 1167716870
## 6 Storm Surge/Ti… 870042000 0 870042000
## 7 Flood 649031930 4905300 653937230
## 8 Thunderstorm W… 412498090 29056210 441554300
## 9 Hail 226989630 142860000 369849630
## 10 High Wind 127967600 112595200 240562800
economic_plot <- economic_impact %>%
arrange(desc(Total_Economic_Damage)) %>%
slice_head(n = 10) %>%
ggplot(aes(x = reorder(EVENT_TYPE, Total_Economic_Damage),
y = Total_Economic_Damage / 1000000000)) +
geom_col(fill = "darkorange") +
geom_text(
aes(label = paste0("$", round(Total_Economic_Damage / 1000000000, 2), "B")),
hjust = -0.2
) +
coord_flip(clip = "off") +
scale_y_continuous(expand = expansion(mult = c(0, 0.18))) +
labs(
title = "Top 10 Storm Events Causing Economic Damage",
x = "Storm Event Type",
y = "Total Economic Damage (Billions of Dollars)",
caption = "Source: NOAA Storm Events Database, 2024"
) +
theme_minimal()
print(economic_plot)
Figure 4. The ten storm event types with the greatest combined property and crop damage in 2024, measured in billions of dollars.
The analysis of the 2024 NOAA Storm Events Database shows that different types of storms create different challenges for communities across the United States. Tornadoes caused the most injuries, while excessive heat caused the most fatalities. Thunderstorm wind was the most frequently reported event type overall, although the most common storm types varied geographically and seasonally.
The economic analysis showed that hurricanes caused the greatest financial damage, followed by flash floods and tropical storms. These findings highlight the importance of preparation for different types of severe weather based on geographic location and seasonal patterns.
Overall, the results can help government officials and emergency management organizations identify important weather-related risks, prioritize preparation efforts, and better allocate financial resources to protect public safety and property.