Impact of weather events on people and properties

Synonpsys

There are weather events which can have a huge impact on people and properties. To understand better the impact on the people health and damages done to properties an analyses of which type weather events have the highest impact will help to take the proper prevention activities.

The analyses of the data collected by US National Weather Service will provide a summary of the top ten weather events with the highest impact on people health and properties. The health impact is related to the number of people which suffered injuries and the number of fatalities because of the weather. The damage done by sever weather events is grouped in properties damage and crop damage.

Based on the type of events the proper resources and activities can take place to minimize the impact and consequences of the weather event.

Data Processing

Storm Data used in the analyses is an official publication of the National Oceanic and Atmospheric Administration Data used in the analyses was collected by US National Weather Service. The data was provided in afile in a CSV format at the next link https://d396qusza40orc.cloudfront.net/repdata%2Fdata%2FStormData.csv.bz2. The file was downloaded on 20th May, 2014 and uncompresssed using Winrar.

R statistics was used to analyze the data and summarize the weather events which are most harmfull to health and produce the most damages.

Data loading

• Setting the working directory

  setwd("D:\\cursuri\\ReproducibleResearch\\PeerAssessment2")
  

• Loading the data in a data frame, sd.

  sd <- read.csv("repdata-data-StormData.csv")
  

• Data transformation

  Change all the events type to upper case

  sd$EVTYPE <- toupper(sd$EVTYPE)
  

  Eliminate the punctuation from the events type name

  sd$EVTYPE <- gsub("\\.", "", sd$EVTYPE)
  

Summarize/grouping data per event type

There will be four main type of summaries based on number of fatalities, injuries and the amount of properties and crop damages. From the summaries will be drop the weather events weith zero sum.

• Summary of weather events based on the number of fatalities

  Fatalities <- aggregate(FATALITIES ~ EVTYPE, data = sd, FUN = sum, na.rm = TRUE)
  Fatalities <- subset(Fatalities, Fatalities$FATALITIES > 0)
  

• Find the first top 10 weather events based on fatalities

  Fatalities10 <- Fatalities[order(Fatalities$FATALITIES, decreasing = TRUE), 
    ]
  Fatalities10 <- Fatalities10[1:10, ]
  

• Summary of weather events based on the number of injuries

  Injuries <- aggregate(INJURIES ~ EVTYPE, data = sd, FUN = sum, na.rm = TRUE)
  Injuries <- subset(Injuries, Injuries$INJURIES > 0)
  

• Find the first top 10 weather events based on injuries

  Injuries10 <- Injuries[order(Injuries$INJURIES, decreasing = TRUE), ]
  Injuries10 <- Injuries10[1:10, ]
  

• Merge the top 10 fatalities and injuries by weather event type

  HealthEvents10 <- merge(Fatalities10, Injuries10, by.x = "EVTYPE", all = TRUE)
  HealthEvents10[is.na(HealthEvents10)] <- 0
  HealthEvents10$EVTYPE <- as.factor(HealthEvents10$EVTYPE)
  

• Summary of weather events based on properties damage

  PropertyDamage <- aggregate(PROPDMG ~ EVTYPE, data = sd, FUN = sum, na.rm = TRUE)
  PropertyDamage <- subset(PropertyDamage, PropertyDamage$PROPDMG > 0)
  

• Find the first top 10 weather events based on property damage

  PropertyDamage10 <- PropertyDamage[order(PropertyDamage$PROPDMG, decreasing = TRUE), 
    ]
  PropertyDamage10 <- PropertyDamage10[1:10, ]
  

• Summary of weather events based on crop damage

  CropDamage <- aggregate(CROPDMG ~ EVTYPE, data = sd, FUN = sum, na.rm = TRUE)
  CropDamage <- subset(CropDamage, CropDamage$CROPDMG > 0)
  

• Find the first top 10 weather events based on crop damage

  CropDamage10 <- CropDamage[order(CropDamage$CROPDMG, decreasing = TRUE), ]
  CropDamage10 <- CropDamage10[1:10, ]
  

• Merge the top 10 property and crop damage by weather event type

  Damage10 <- merge(PropertyDamage10, CropDamage10, by.x = "EVTYPE", all = TRUE)
  Damage10[is.na(Damage10)] <- 0
  Damage10$EVTYPE <- as.factor(Damage10$EVTYPE)
  

Results

Summary impact of extreme weather events on people health

• Based on the data analysed the weather event with the most impact on people, the higher number of fatalities happend because of tornado and the number of fatalities is 5633.

• The weather event with the highest number of injuries is tornado with the total number of injuries of 9.1346 × 10⁴.

• Here is a summary with the top 10 weather events based on number of fatalities and injuries. The data is order by fatalities and injuries.

  HealthEvents10[order(HealthEvents10$FATALITIES, HealthEvents10$INJURIES, decreasing = TRUE), 
    ]
  

  ##               EVTYPE FATALITIES INJURIES
  ## 12           TORNADO       5633    91346
  ## 2     EXCESSIVE HEAT       1903     6525
  ## 3        FLASH FLOOD        978     1777
  ## 6               HEAT        937     2100
  ## 9          LIGHTNING        817     5230
  ## 13         TSTM WIND        504     6957
  ## 4              FLOOD        470     6789
  ## 10       RIP CURRENT        368        0
  ## 7          HIGH WIND        248        0
  ## 1          AVALANCHE        224        0
  ## 8          ICE STORM          0     1975
  ## 11 THUNDERSTORM WIND          0     1488
  ## 5               HAIL          0     1361
  

• A bar plot of the data presented above is draw for an easy analyses

  library(lattice)
  barchart(EVTYPE ~ FATALITIES + INJURIES, HealthEvents10, main = "Health events", 
    xlab = "Health events", auto.key = list(space = "top", columns = 2), scales = list(x = list(log = TRUE)))
  

  

Summary impact of extreme weather events on people properties and crop

• Based on the data analysed the weather event with the most impact properties, the highest amount of damages, happend because of tornado and the amount of damage is 3.2123 × 10⁶.

  • The weather event with the highest number of injuries is hail with the total number of injuries of 5.796 × 10⁵.
  • Here is a summary with the top 10 weather events based on property and crop damage. The data is order by property and crop damage.

  Damage10[order(Damage10$PROPDMG, Damage10$CROPDMG, decreasing = TRUE), ]
  

  ##                EVTYPE PROPDMG CROPDMG
  ## 10            TORNADO 3212258  100019
  ## 2         FLASH FLOOD 1420125  179200
  ## 11          TSTM WIND 1335996  109203
  ## 3               FLOOD  899938  168038
  ## 8   THUNDERSTORM WIND  876844   66794
  ## 4                HAIL  688693  579596
  ## 7           LIGHTNING  603352       0
  ## 9  THUNDERSTORM WINDS  446298   18685
  ## 6           HIGH WIND  324732   17283
  ## 12       WINTER STORM  132721       0
  ## 1             DROUGHT       0   33899
  ## 5          HEAVY RAIN       0   11123
  

  • A bar plot of the data presented above is draw for an easy analyses

  library(lattice)
  barchart(EVTYPE ~ PROPDMG + CROPDMG, Damage10, main = "Weather events damagess", 
    xlab = "Damage amount (dollars)", auto.key = list(space = "top", columns = 2, 
        text = c("Property damage", "Crop damage")), scales = list(x = list(log = TRUE)))