Dataset first overview

Please note that I saved the excel file as CSV on my computer.

After loading the file and having a quick view at the data, four small issues occure:

• each column name (except the first one) start by “X”

• last column named “X” is empty

• last row is empty

• columns are not variables

library(ggplot2)
library(dplyr)
library(tidyr)
library(ggrepel)

data.source <- read.csv("C:/Users/marc/Desktop/Data/160817_plane crash/indicator_air accident killed.csv", sep = ";", header= TRUE, stringsAsFactors = FALSE)
data.source <- tbl_df(data.source)
dim(data.source)  # number of rows and columns

## [1] 134  41

colnames(data.source) 

##  [1] "Air.accidents.killed" "X1970"                "X1971"               
##  [4] "X1972"                "X1973"                "X1974"               
##  [7] "X1975"                "X1976"                "X1977"               
## [10] "X1978"                "X1979"                "X1980"               
## [13] "X1981"                "X1982"                "X1983"               
## [16] "X1984"                "X1985"                "X1986"               
## [19] "X1987"                "X1988"                "X1989"               
## [22] "X1990"                "X1991"                "X1992"               
## [25] "X1993"                "X1994"                "X1995"               
## [28] "X1996"                "X1997"                "X1998"               
## [31] "X1999"                "X2000"                "X2001"               
## [34] "X2002"                "X2003"                "X2004"               
## [37] "X2005"                "X2006"                "X2007"               
## [40] "X2008"                "X"

head(data.source[,1:5], 5) #first rows and columns

## # A tibble: 5 x 5
##   Air.accidents.killed X1970 X1971 X1972 X1973
##                  <chr> <int> <int> <int> <int>
## 1          Afghanistan     0     0     0     0
## 2              Algeria     0     0     0     0
## 3               Angola     0     0     0     0
## 4            Argentina     0     0     0     0
## 5              Armenia     0     0     0     0

tail(data.source[,1:5],5)

## # A tibble: 5 x 5
##   Air.accidents.killed X1970 X1971 X1972 X1973
##                  <chr> <int> <int> <int> <int>
## 1         West Germany     0     0     0     0
## 2          Yemen, Rep.     0     0     0     0
## 3           Yugoslavia     0     0     0     0
## 4               Zambia     0     0     0     0
## 5                         NA    NA    NA    NA

Data cleaning

#Removing last empty row and column
data.cleaned <- data.source[-134,-41]

#Moving columns into Rows
data.cleaned <- gather(data.cleaned, year, people.killed, -Air.accidents.killed)
head(data.cleaned,5)

## # A tibble: 5 x 3
##   Air.accidents.killed  year people.killed
##                  <chr> <chr>         <int>
## 1          Afghanistan X1970             0
## 2              Algeria X1970             0
## 3               Angola X1970             0
## 4            Argentina X1970             0
## 5              Armenia X1970             0

colnames(data.cleaned) <- c("country", "year", "people.killed")

#remove the "X" in front of each year
data.cleaned$year <- gsub("X", "", data.cleaned$year)
data.cleaned$year <- as.numeric(data.cleaned$year)

Number of people killed per year

This dataset is from 1970 to 2008

group.year <- group_by(data.cleaned, year)
ppl.killed.per.year <- summarize(group.year, nb.death = sum(people.killed)) 

ggplot(ppl.killed.per.year, aes(x=year, y= nb.death, group=1, label=nb.death)) +
      geom_point(size=5, color= "#008080")+
      geom_line(size= 1.2, color= "#008080")+
      theme_bw()+
      geom_label_repel(color = "black", 
                       box.padding = unit(0.25, "lines"))+
      ylab("Number of people killed") +
      ggtitle("Number of people killed by air accidents per year for all countries")+
      scale_x_continuous(breaks= seq(1970,2008,3))

1996 stands out as being a tragic year. We can check to see in which countries were the main incidents this year.

year.1996 <- filter(data.cleaned, year == 1996 & people.killed != 0) %>%
      arrange(desc(people.killed))      

knitr::kable(head(year.1996,10))

Top 15 countries with the most people killed in air accidents

killed.country <- group_by(data.cleaned, country) %>% 
      summarize(death = sum(people.killed)) %>%
      arrange(desc(death))

head15 <- head(killed.country,15)


ggplot(head15, aes(x=country, y=death)) +
      geom_bar(stat = "identity", fill= "#008080") +
      coord_flip()+
      theme_bw()

Top 15 countries with the least people killed in air accidents

bottom15 <- tail(killed.country,15)

ggplot(bottom15, aes(x=country, y=death)) +
      geom_bar(stat = "identity", fill= "#008080") +
      coord_flip()+
      theme_bw()

Worst incident(s) for a given year and country

A small research on Wikipedia indicates that the worst incident in the canari island is related to “The Tenerife airport disaster [which was] was a fatal runway collision between two Boeing 747s on Sunday, March 27, 1977”

knitr::kable(head(arrange(data.cleaned,desc(people.killed)),5))