############################################################
## 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 6  ###
###################

###############
## Histogram ##
###############

# Read in GSS data
gss <- read.csv(file='https://raw.githubusercontent.com/artofstat/data/master/Chapter2/TVhours.csv')

# Create Basic Histogram
hist(gss$tvhours, xlab = 'Hours of Watching TV', main = 'Histogram of the Numbers of Hours Watching TV per Day')

# For more fine tuning, it is better to use the ggplot2 library
# If you haven't installed it already, first type: install.packages(ggplot2)
library(ggplot2)

# Basic Frequency (Count) Plot
ggplot(gss, aes(x=tvhours)) + 
  geom_histogram(binwidth=1, color = 'black', fill = 'orchid') +
  labs(x = 'Hours of Watching TV', y='Frequency',  
       title = 'Histogram of the Numbers of Hours Watching TV per Day', 
       subtitle = 'Source: General Social Survey 2012') +
  theme_bw()

# Basic Relative Frequency (Percent) Plot
ggplot(gss, aes(x = tvhours, y = 100 * (..count.. / sum(..count..)))) + 
  geom_histogram(binwidth = 1, color = 'black', fill = 'orchid') +
  labs(x = 'Hours of Watching TV', y = 'Percent (%)',  
       title = 'Histogram of the Numbers of Hours Watching TV per Day', 
       subtitle = 'Source: General Social Survey 2012') +
  theme_bw()

# More fine-tuning such as labels on x-axis 
ggplot(gss, aes(x = tvhours)) + 
  geom_histogram(center = 0, binwidth = 1, color = 'black', fill = 'orchid') +
  labs(x = 'Hours of Watching TV', y = 'Frequency',  
       title = 'Histogram of the Numbers of Hours Watching TV per Day', 
       subtitle = 'Source: General Social Survey 2012') +
  theme_bw() +
  scale_y_continuous(limits = c(0,350), breaks = seq(0,350,50), expand = c(0,0)) +
  scale_x_continuous(breaks = seq(0,25,1)) +
  theme(panel.grid.major.x = element_blank(), panel.grid.minor.x = element_blank())

# with percentages rather than counts
ggplot(gss, aes(x = tvhours, y = 100 * (..count.. / sum(..count..)))) + 
  geom_histogram(center = 0, binwidth = 1, color = 'black', fill = 'orchid') +
  labs(x = 'Hours of Watching TV', y = 'Percent (%)',  
       title = 'Histogram of the Numbers of Hours Watching TV per Day', 
       subtitle = 'Source: General Social Survey 2012') +
  theme_bw() +
  scale_y_continuous(limits = c(0,26), breaks = seq(0,25,5), expand = c(0,0)) +
  scale_x_continuous(breaks = seq(0,25,1)) +
  theme(panel.grid.major.x = element_blank(), panel.grid.minor.x = element_blank())