Severe Weather Events: Health and Economic impact accros the US

Synopsis

The data analyzed in this document compiles observations on multiple weather events and multiple features of these events including location, duration, etc. This analysis is centered on the events type, their impact on citizens life( injuries and deaths, hereby referred as “Health”) and the economic damage caused between 1990 and 2011. This document will recapitulate the data analysis and present the main results.

Data Processing

The following section describes in detail how the data was processed.

#download file
download.file("https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2", "repdata_data_StormData.csv.bz2", mode = "wb")

# Unzip the file
unzip("repdata_data_StormData.csv.bz2")

#read the file
StormData <- read_csv("repdata_data_StormData.csv")

## Rows: 902297 Columns: 37
## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr (18): BGN_DATE, BGN_TIME, TIME_ZONE, COUNTYNAME, STATE, EVTYPE, BGN_AZI,...
## dbl (18): STATE__, COUNTY, BGN_RANGE, COUNTY_END, END_RANGE, LENGTH, WIDTH, ...
## lgl  (1): COUNTYENDN
## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

The first manipulation of the dataset was the fusion of the variables related health (fatalities and injuries) into a single variable named “Health”, and the fusion of variables related to economic damages (property and crops) into a single variable named “Ec”. I also made sure to transform the date to POSIXlt format and make a column with only the year in it to simplify analysis.

#create new columns to add the variables related to human health and to economic impact
StormData$Health <- StormData$FATALITIES+StormData$INJURIES
StormData$Ec <- StormData$PROPDMG+StormData$CROPDMG

#create new columns to keep track of the dates
StormData$DATE <- as.POSIXlt(StormData$BGN_DATE, format = "%m/%d/%Y %H:%M:%S")
StormData$Year <- as.numeric(format(StormData$DATE, "%Y"))

After these manipulations, I summarized the data by year and event, keeping the total sum of damages caused to represent the damages to the population’s health and to the economy. Observations outside of the time-frame and observations without any kind of damage were excluded from the database to simplify analysis.

#Create new dataframe to summarise the information by year and event
df <- StormData %>%
  group_by(Year, EVTYPE) %>%
  summarise(
    total_fatalities = sum(FATALITIES),
    total_Injuries = sum(INJURIES),
    Health = sum(Health),
    total_Prop = sum(PROPDMG),
    total_Crop = sum(CROPDMG),
    Ec = sum(Ec))

## `summarise()` has grouped output by 'Year'. You can override using the
## `.groups` argument.

#get rid of observations without any effect on health and the economy, limit observations to after 1990

df <- subset(df, df$Health>0 & df$Ec>0 & df$Year>1990)

Given that the original database contained almost a thousand unique event types, making analysis almost impossible, I created a second “event” variable to limit the number of categories. In this second variable, I grouped events with similar fundamental causes into a single category. For example, hail, rain and snow events were all encompassed by the category “precipitations”. In total, 10 categories were kept: Cold/Winter, Heat, Flood, Hurricane, Tornado, Precipitations, Storm, Wind, Wild Fires and Other, each created according to the criteria shown in the next chunk of code.

#Simplify the event types

df <- df %>% mutate(Event = case_when(
      grepl("TORNADO", EVTYPE, ignore.case = TRUE) ~ "TORNADO",
      grepl("HURRICANE", EVTYPE, ignore.case = TRUE) ~ "HURRICANE",
      grepl("HEAT", EVTYPE, ignore.case = TRUE) ~ "HEAT",
      grepl("WIND", EVTYPE, ignore.case = TRUE) ~ "WIND",
      grepl("FLOOD", EVTYPE, ignore.case = TRUE) ~ "FLOOD",
      str_detect(EVTYPE, regex("STORM|LIGHTNING", ignore_case = TRUE)) ~ "STORM",
      str_detect(EVTYPE, regex("COLD|FREEZING|FREEZE|ICE|ICY|FROST|WINTER", ignore_case = TRUE)) ~ "COLD/WINTER",
      str_detect(EVTYPE, regex("HAIL|RAIN|SNOW|THUNDERSNOW|BLIZZARD", ignore_case = TRUE)) ~ "PRECIPITATIONS",
      str_detect(EVTYPE, regex("FIRE", ignore_case = TRUE)) ~ "WILD FIRES",
      TRUE ~ "OTHER"  # Default case if none of the conditions match
    ))

Further, I created another dataframe with the informations grouped by event in order to see the aggregated damages over the entire period.

Having finished the data processing, my next step was to visualize the data using ggplot to illustrate the damages caused throughout the years by event type. One visualization was made for health-related damage and another for economic damage.

# Create a ggplot for overall health damage
ggplot(df, aes(x = Year, y = Health, color = Event)) +
  geom_line() +
  labs(title = "Injuries and Deaths Over the Years",
       x = "Year",
       y = "Number of Injuries and Deaths") +
  facet_wrap(~ Event, scales = "free_y", ncol = 2) +
  theme_minimal()

# Create a ggplot for overall economic damage

 ggplot(df, aes(x = Year, y = Ec, color = Event)) +
  geom_line() +
  labs(title = "Economic Impact Over the Years",
       x = "Year",
       y = "Economic Impact (in") +
  facet_wrap(~ Event, scales = "free_y", ncol = 2) +
  theme_minimal()

Results

The analysis shows that punctual events such as floods and tornadoes have historically caused the greater damage to health with upwards of 6 thousands victims in a single year Economic damage is headed by the event type “wind”, followed by floods and tornadoes with maximums of 200 000 millions in costs in a single year

However, these extreme historical events do not seem to represent the overall damages caused in this period. While having caused little damage in a single year, storms caused more than 10 thousand victims in aggregate, a number close to the one produced by floods. Similarly, while tornadoes caused the most casualties in a single year, in total heat caused more deaths over this period. The following table summarizes the damages by event.

Summary of Health and Economic Damages per Event Type
Event	Fatalities	Annual_Fatalities	Injuries	Annual_Injuries	Total_Health	Annual_Health	Property_Damage	Annual_Property_Damage	Crop_Damage	Annual_Crop_Damage	Total_Economic_Damage	Annual_Economic_Damage
TORNADO	1727	82	25558	1217	27285	1299	1520700	72414	100021	4763	1620721	77177
STORM	1232	59	9429	449	10661	508	859680	40937	15616	744	875295	41681
FLOOD	1523	73	8601	410	10124	482	2393708	113986	361470	17213	2755178	131199
WIND	1065	51	8159	389	9224	439	3126404	148876	222922	10615	3349327	159492
HEAT	2244	107	5716	272	7960	379	3233	154	1473	70	4706	224
PRECIPITATIONS	345	16	3112	148	3457	165	885041	42145	594573	28313	1479614	70458
OTHER	694	33	2214	105	2908	138	73619	3506	12483	594	86103	4100
WILD FIRES	90	4	1606	76	1696	81	122912	5853	8549	407	131462	6260
HURRICANE	135	6	1326	63	1461	70	20936	997	9783	466	30719	1463
COLD/WINTER	252	12	882	42	1134	54	36720	1749	6206	296	42926	2044

StormData

DaniGCFH

2024-01-19

Severe Weather Events: Health and Economic impact accros the US

Synopsis

Data Processing

Results