NOAA Storm Data Analysis: Identifying the Most Harmful Weather Events to Public Health and Their Economic Impact

Introducction: # This analysis examined the U.S. National Weather Service (NWS) Storm Database to : 1 Most Harmful Events to Population Health; 2 Events with Greatest Economic Consequences

Data Processing

data loading

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)
#dir("repdata_data_StormData")
dat1 <- read.csv("repdata_data_StormData/repdata_data_StormData.csv")

data analysis

1. To detremin which types of events (as indicated in the EVTYPE) are most harmful with respect to population health

harm_summary <- tapply(dat1$FATALITIES, dat1$EVTYPE, sum, na.rm = TRUE) + 
                tapply(dat1$INJURIES, dat1$EVTYPE, sum, na.rm = TRUE) %>%
                as.data.frame() %>%
                rename(total_harm = ".")


harm_summary <- harm_summary %>% 
                arrange(desc(total_harm)) %>% as.data.frame()
harm_summary$EVTYPE <- rownames(harm_summary)
# View the top 5 most harmful events

Justification: The raw data contains separate columns for FATALITIES and INJURIES, requiring aggregation to accurately assess total population health impact. By summing these metric

result 1:

ggplot(harm_summary[1:5, ], aes(x = reorder(EVTYPE, -total_harm), y = total_harm)) +
  geom_col(fill = "blue") +
  geom_text(aes(label = total_harm), vjust = -0.5) +
  labs(title = "Top 5 Most Harmful Weather Events to Population Health",
       x = "Event Type",
       y = "Total Harm (Fatalities + Injuries)") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) + theme_classic()

2. To determine which types of events cause the greatest economic losses

Damage Value Conversion: Description: Transformation: Combined PROPDMG/CROPDMG with their EXP multipliers (K=1,000, M=1,000,000, B=1,000,000,000)

# Calculate total economic impact by event type
economic_impact <- dat1 %>%
  mutate(
    # Convert damage amounts to actual values (using EXP columns)
    prop_dmg_value = PROPDMG * case_when(
      PROPDMGEXP == "K" ~ 1e3,
      PROPDMGEXP == "M" ~ 1e6,
      PROPDMGEXP == "B" ~ 1e9,
      TRUE ~ 1
    ),
    crop_dmg_value = CROPDMG * case_when(
      CROPDMGEXP == "K" ~ 1e3,
      CROPDMGEXP == "M" ~ 1e6,
      CROPDMGEXP == "B" ~ 1e9,
      TRUE ~ 1
    )
  )
economic_impact <- tapply(economic_impact$prop_dmg_value, economic_impact$EVTYPE, sum, na.rm = TRUE) + 
                tapply(economic_impact$crop_dmg_value,economic_impact$EVTYPE, sum, na.rm = TRUE) %>%
                as.data.frame() %>%
                rename(total_damage = ".") %>% as.data.frame()
economic_impact$EVTYPE <- rownames(economic_impact)

economic_impact <- economic_impact  %>% arrange(desc(total_damage))

Justification: Raw data stores damage values separately from their units, making aggregation impossible without conversion to consistent units

result 2:

ggplot(economic_impact[1:5, ], aes(x = reorder(EVTYPE, -total_damage), y = total_damage)) +
  geom_col(fill = "blue") +
  geom_text(aes(label = total_damage), vjust = -0.5) +
  labs(title = "Top 5 Most Harmful Weather Events to Total Damage",
       x = "Event Type",
       y = "Total Damage (Crop Damage + Property Damage)") +
  theme(axis.text.x = element_text(angle = 45, hjust = 1)) + theme_classic()

Synopsis:

This analysis of NOAA Storm Data reveals tornadoes as the most dangerous events to population health, causing the majority of weather-related fatalities and injuries. For economic damage, floods and hurricanes are the costliest, with billions in property and crop losses. Data was standardized to compare impacts consistently, using transformations like damage value conversion (e.g., “1.5B” → $1.5 billion) and event-type normalization. The findings highlight tornado preparedness and flood-resistant infrastructure as critical priorities for public safety and economic stability.