#############################################################
## R code to reproduce statistical analysis in the textbook:
## Agresti, Franklin, Klingenberg
## Statistics: The Art & Science of Learning from Data
## 5th Edition, Pearson 2021
## Web: ArtofStat.com
## Copyright: Bernhard Klingenberg
############################################################

###################
###  Chapter 2  ###
###  Example 3  ###
###################

###############################
## Create Pie and Bar Charts ##
###############################

# Create dataset:
state <- c('Florida', 'Hawaii', 'South \n Carolina', 'California', 'North \n Carolina', 'Texas', 'Other')
frequency <- c(203, 51, 34, 33, 23, 16, 27)
# Create basic piechart:
pie(frequency, labels=state, col=rainbow(7), main='Pie Chart of Shark Attacks')

# Create basic bar graph showing counts:
barplot(frequency, names.arg = state, cex.names=0.8, col=rainbow(7), xlab='State', ylab='Frequency', main='Bar Chart of Shark Attacks')

# Create basic bar graph showing percentages:
percent <- 100 * (frequency / sum(frequency))
barplot(percent, names.arg = state, cex.names=0.8, col=rainbow(7), xlab='State', ylab='Percent (%)', main='Bar Chart of Shark Attacks')

# A better-looking bar graph can be obtained with the ggplot2 library
# To install the ggplot2 library, use `install.packages('ggplot2')`
library(ggplot2)
plotdata = data.frame(state, frequency)
ggplot(data = plotdata, 
       aes(x = state, 
           y = frequency,
           fill = state)) + 
  geom_col(show.legend = FALSE) +
  labs(title = 'Bar Chart of Shark Attacks')

# Or use the `geom_bar` function with `stat = identity`
ggplot(data = plotdata, 
       aes(x = state, 
           y = frequency,
           fill = state)) + 
  geom_bar(stat = 'identity',
           show.legend = FALSE) + 
  labs(title = 'Bar Chart of Shark Attacks')

# To manually sort in decreasing order but with `Other` as the last category
# you can use the mutate() function from the dplyr package along with the fct_relvel() function from the forcats package
# To install both packages, use `install.packages('dplyr')` and `install.packages('forcats')`
library(dplyr)
library(forcats)
plotdata1 = plotdata %>% 
  mutate(state = fct_relevel(state, c('Florida', 'Hawaii', 'South \n Carolina', 
                                      'California', 'North \n Carolina', 'Texas', 'Other')))
ggplot(data = plotdata1, 
       aes(x = state, 
           y = frequency,
           fill = state)) + 
  geom_col(show.legend = FALSE) + 
  labs(title = 'Bar Chart of Shark Attacks') +
  theme_classic()

# Note that the ggplot2, dplyr, and forcats packages are all conveniently included within the tidyverse package
# So you can instead use install.packages('tidyverse') and then use the library(tidyverse) function