The analisys of the most harmful weather events in US

Synopsis

The analysis is based on the U.S. National Oceanic and Atmospheric Administration's (NOAA) storm database. This database tracks characteristics of major storms and weather events in the United States, including when and where they occur, as well as estimates of any fatalities, injuries, and property damage. In this work we explore the data for the purpose to answer two questions:

  1. Across the United States, which types of events (as indicated in the EVTYPE variable) are most harmful with respect to population health? In this section we assess the number of fatalities and injuries with respect to types of events.

  2. Across the United States, which types of events have the greatest economic consequences? In this section we assess the amounts of property damage with respect to types of events.

The main method which we use in the analysis is frequency analysis. The results shows that we can pick out “Top 10” of the most harmful types of events with respect to population health (namely fatalities and injures) and property damages (see section Results).

Data Processing

The analisys includes the following steps.

  1. The dataset reading:
data = read.csv("C:/Users/User/Dropbox/reproducible research/assigment2/repdata_data_StormData.csv.bz2")
  1. Exploring population health in respect with types of recorded events. There are two variables in the data set (FATALITIES and INJURIES) which are represented the information about damages of public health.

The first step is to calculate the total number of fatalities across the types of events.

sum.fatal.by.type = by(data$FATALITIES, data$EVTYPE, sum, na.rm = T)

Next we find the types of events in which non-zero fatalities are observed and sort them in increasing order.

sum.fatal.nonzero = sort(sum.fatal.by.type[sum.fatal.by.type > 0])

We consider as the most harmful the last 10 types of events in the variational series.

tail(sum.fatal.nonzero, 10)
## data$EVTYPE
##      AVALANCHE      HIGH WIND    RIP CURRENT          FLOOD      TSTM WIND 
##            224            248            368            470            504 
##      LIGHTNING           HEAT    FLASH FLOOD EXCESSIVE HEAT        TORNADO 
##            816            937            978           1903           5633

Among this list we can emphasize


Error in names(tail(sum.fatal.nonzero), 1) : 
  2 аргумента переданы 'names', а требуется 1

and EXCESSIVE HEAT as the most dangerous events in sense of mortality.

Next we examine the number of injures in the same way as fatalities.

sum.inj.by.type = by(data$INJURIES, data$EVTYPE, sum, na.rm = T)
sum.inj.nonzero = sort(sum.inj.by.type[sum.inj.by.type > 0])
tail(sum.inj.nonzero, 10)
## data$EVTYPE
##              HAIL THUNDERSTORM WIND       FLASH FLOOD         ICE STORM 
##              1361              1488              1777              1975 
##              HEAT         LIGHTNING    EXCESSIVE HEAT             FLOOD 
##              2100              5230              6525              6789 
##         TSTM WIND           TORNADO 
##              6957             91346

In this case we can mark out TORNADO as the most harmful type of events.

  1. Extracting the types of events which have the greatest economic consequences

We examine the amounts of property damage in the same way as fatalities and injuries. We consider the amounts of damage more $100000 as outstanding values

sum.propdmg.by.type = by(data$PROPDMG, data$EVTYPE, sum, na.rm = T)
sum.propdmg.nonzero = sort(sum.propdmg.by.type[sum.propdmg.by.type > 0])

sum.propdmg.nonzero[sum.propdmg.nonzero > 1e+05]
## data$EVTYPE
##         HEAVY SNOW       WINTER STORM          HIGH WIND 
##             122252             132721             324732 
## THUNDERSTORM WINDS          LIGHTNING               HAIL 
##             446293             603352             688693 
##  THUNDERSTORM WIND              FLOOD          TSTM WIND 
##             876844             899938            1335966 
##        FLASH FLOOD            TORNADO 
##            1420125            3212258

Results

The most harmful types of events with respect to population health

We used the numbers of fatalities and injuries as the main measurements of harmfulness of the events. The distributions of the two variables are represented in the figure 1.

plot of chunk unnamed-chunk-7

Figure 1. Distribution of total numbers of fatalities (left panel) and injures (right panel) in different types of events.

In the fig. 1 we can see two obvious outliers in fatalities. These are TORNADO (the most dangerous) and EXCESSIVE HEAT (the second).

In addition we can pick out “Top 10” of the most harmful in respect of mortality events:

## data$EVTYPE
##      AVALANCHE      HIGH WIND    RIP CURRENT          FLOOD      TSTM WIND 
##            224            248            368            470            504 
##      LIGHTNING           HEAT    FLASH FLOOD EXCESSIVE HEAT        TORNADO 
##            816            937            978           1903           5633

In the right panel of fig.1 there is one remarkable outlier, this is TORNADO. The “Top 10 most harmful types of events” in this case is the follow:

## data$EVTYPE
##              HAIL THUNDERSTORM WIND       FLASH FLOOD         ICE STORM 
##              1361              1488              1777              1975 
##              HEAT         LIGHTNING    EXCESSIVE HEAT             FLOOD 
##              2100              5230              6525              6789 
##         TSTM WIND           TORNADO 
##              6957             91346

Types of events with the greatest economic consequences

The distribution of amounts of property damages in different types of events is presented in figure 2.

plot of chunk unnamed-chunk-10

Figure 2. Distribution of total amounts of property damages in different types of events.

There is one outlier in the figure 2 - 3.2123 × 106.

If we consider the amounts of damage more $100000 as remarkable, there are 11 the most destructive types of events:

## data$EVTYPE
##         HEAVY SNOW       WINTER STORM          HIGH WIND 
##             122252             132721             324732 
## THUNDERSTORM WINDS          LIGHTNING               HAIL 
##             446293             603352             688693 
##  THUNDERSTORM WIND              FLOOD          TSTM WIND 
##             876844             899938            1335966 
##        FLASH FLOOD            TORNADO 
##            1420125            3212258

Summary

The most dangerous with respect to as population health as economic consequences are tornados. Beside this type of events we can pick out the others, the lists are presented in Results section.