Severe Weather Events: Health and Economic Impacts in the United States

Synopsis

Storms and severe weather events cause significant public health and economic damage across the United States. This analysis explores the NOAA Storm Database (1950–2011) to identify which event types are most harmful to population health (measured by fatalities and injuries) and which have the greatest economic consequences (measured by property and crop damage).

Tornadoes are by far the most harmful to human health, accounting for the vast majority of fatalities and injuries. For economic impact, floods cause the greatest total damage, followed by hurricanes/typhoons and tornadoes. The analysis processes the raw compressed CSV file directly in R and presents results with two figures (one combined health figure and one economic figure).

Data Processing

library(R.utils)
library(data.table)
library(dplyr)
library(ggplot2)
library(tidyr)

if (!file.exists("repdata_data_StormData.csv.bz2")) {
  download.file("https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2",
                "repdata_data_StormData.csv.bz2", method = "curl")
}

if (!file.exists("repdata_data_StormData.csv")) {
  bunzip2("repdata_data_StormData.csv.bz2", "repdata_data_StormData.csv", remove = FALSE)
}

storm <- fread("repdata_data_StormData.csv", fill = TRUE, header = TRUE)
dim(storm)

## [1] 902297     37

head(storm, 3)

##    STATE__          BGN_DATE BGN_TIME TIME_ZONE COUNTY COUNTYNAME  STATE
##     <char>            <char>   <char>    <char> <char>     <char> <char>
## 1:    1.00 4/18/1950 0:00:00     0130       CST  97.00     MOBILE     AL
## 2:    1.00 4/18/1950 0:00:00     0145       CST   3.00    BALDWIN     AL
## 3:    1.00 2/20/1951 0:00:00     1600       CST  57.00    FAYETTE     AL
##     EVTYPE BGN_RANGE BGN_AZI BGN_LOCATI END_DATE END_TIME COUNTY_END COUNTYENDN
##     <char>    <char>  <char>     <char>   <char>   <char>     <char>     <char>
## 1: TORNADO      0.00                                            0.00           
## 2: TORNADO      0.00                                            0.00           
## 3: TORNADO      0.00                                            0.00           
##    END_RANGE END_AZI END_LOCATI LENGTH  WIDTH      F   MAG FATALITIES INJURIES
##       <char>  <char>     <char> <char> <char> <char> <num>      <num>    <num>
## 1:      0.00                     14.00 100.00      3     0          0       15
## 2:      0.00                      2.00 150.00      2     0          0        0
## 3:      0.00                      0.10 123.00      2     0          0        2
##    PROPDMG PROPDMGEXP CROPDMG CROPDMGEXP    WFO STATEOFFIC ZONENAMES LATITUDE
##      <num>     <char>   <num>     <char> <char>     <char>    <char>    <num>
## 1:    25.0          K       0                                            3040
## 2:     2.5          K       0                                            3042
## 3:    25.0          K       0                                            3340
##    LONGITUDE LATITUDE_E LONGITUDE_ REMARKS REFNUM
##        <num>      <num>      <num>  <char>  <num>
## 1:      8812       3051       8806              1
## 2:      8755          0          0              2
## 3:      8742          0          0              3

Data Transformations and Cleaning

exp_to_num <- function(exp) {
  case_when(
    exp %in% c("K", "k") ~ 1000,
    exp %in% c("M", "m") ~ 1e6,
    exp %in% c("B", "b") ~ 1e9,
    TRUE ~ 1
  )
}

storm_clean <- storm %>%
  mutate(
    PROPDMGEXP_num = exp_to_num(PROPDMGEXP),
    CROPDMGEXP_num = exp_to_num(CROPDMGEXP),
    TotalPropertyDamage = PROPDMG * PROPDMGEXP_num,
    TotalCropDamage = CROPDMG * CROPDMGEXP_num,
    TotalEconomicDamage = TotalPropertyDamage + TotalCropDamage,
    TotalHealthImpact = FATALITIES + INJURIES
  ) %>%
  filter(!is.na(EVTYPE))

health_impact <- storm_clean %>%
  group_by(EVTYPE) %>%
  summarise(
    TotalFatalities = sum(FATALITIES, na.rm = TRUE),
    TotalInjuries = sum(INJURIES, na.rm = TRUE),
    TotalHealth = sum(TotalHealthImpact, na.rm = TRUE)
  ) %>%
  arrange(desc(TotalHealth)) %>%
  slice_head(n = 10)

economic_impact <- storm_clean %>%
  group_by(EVTYPE) %>%
  summarise(
    TotalProperty = sum(TotalPropertyDamage, na.rm = TRUE),
    TotalCrop = sum(TotalCropDamage, na.rm = TRUE),
    TotalEconomic = sum(TotalEconomicDamage, na.rm = TRUE)
  ) %>%
  arrange(desc(TotalEconomic)) %>%
  slice_head(n = 10)

Results

1. Events Most Harmful to Population Health

health_long <- health_impact %>%
  select(EVTYPE, TotalFatalities, TotalInjuries) %>%
  pivot_longer(cols = c(TotalFatalities, TotalInjuries),
               names_to = "Metric", values_to = "Count")

ggplot(health_long, aes(x = reorder(EVTYPE, Count), y = Count, fill = Metric)) +
  geom_col(position = "dodge") +
  coord_flip() +
  labs(title = "Top 10 Events Harmful to Population Health",
       x = "Event Type", y = "Count",
       fill = "Impact") +
  theme_minimal()

Top 10 weather event types by total fatalities and injuries (1950–2011). Tornadoes cause the overwhelming majority of health impacts.

2. Events with the Greatest Economic Consequences

economic_long <- economic_impact %>%
  select(EVTYPE, TotalProperty, TotalCrop) %>%
  pivot_longer(cols = c(TotalProperty, TotalCrop),
               names_to = "DamageType", values_to = "Amount") %>%
  mutate(Amount = Amount / 1e9)

ggplot(economic_long, aes(x = reorder(EVTYPE, Amount), y = Amount, fill = DamageType)) +
  geom_col(position = "stack") +
  coord_flip() +
  labs(title = "Top 10 Events with Greatest Economic Consequences",
       x = "Event Type", y = "Damage (Billions USD)",
       fill = "Damage Type") +
  theme_minimal()

Top 10 weather event types by total economic damage in USD (property + crop damage, 1950–2011). Floods have the greatest overall economic impact.

Summary of Findings

Public Health Priority: Tornadoes cause the most deaths and injuries.
Economic Priority: Floods generate the highest financial losses.

This analysis is fully reproducible from the original raw dataset and uses only two figures.