# Task 1
#Creation of data frame with the data
flo.df<-data.frame(K=c(0,0,1,1), 
                   Flo=c(0,1,0,1),
                   freq=c(31,33,13, 51))

summary(flo.df)

#Conversion of data frame to a contingency table
flo.tab<-xtabs(freq~K+Flo, data=flo.df)
#    Flo
# K    0  1
#   0 31 33
#   1 13 51

# Chisq test
test.res<-chisq.test(flo.tab)
test.res
# 	Pearson's Chi-squared test with Yates'
# 	continuity correction
# 
# data:  chol.tab
# X-squared = 47.289, df = 1, p-value =
# 6.124e-12
test.res$stdres# Standardized residuals
#showing the direction of the association
#    Flo
# K           0         1
#   0  3.349743 -3.349743
#   1 -3.349743  3.349743


# Installing and loading the psych package for function phi
install.packages("psych")
library(psych)

phi(flo.tab)#Phi coefficient - positive indicates
# positive correlation
# [1] 0.3

#Plotting a stacked barplot. geom_col needs to be used
# if the counts are already counted
# Be aware of the content of the aes() - how the x, y and fill are defined 
library(ggplot2)
ggplot(flo.df, aes(x=as.character(K), 
                   y=freq, 
                   fill=as.character(Flo)))+
geom_col()+theme_classic()

#Task 2
tea.df<-data.frame(tea=c(0,0,"g", "g", "b", "b"), 
                   cancer=rep(c(0,1),3), 
                   freq=c(130,55, 357, 65,160,45))
summary(tea.df)

tea.tab<-xtabs(freq~tea+cancer, data=tea.df)
# chisq.test(tea.tab)
# 	Pearson's Chi-squared test
# 
# data:  tea.tab
# X-squared = 16.756, df = 2, p-value
# = 0.0002299

#Instalation and loading of epitools needed for epitab function
# install.packages("epitools")
library(epitools)

# Epitab showing the comparisons of the odds ratios:
# black vs. no tea
# green vs. no tea

epitab(tea.tab)$tab
#    cancer
# tea   0        p0  1        p1 oddsratio
#   0 130 0.2009274 55 0.3333333 1.0000000
#   b 160 0.2472952 45 0.2727273 0.6647727
#   g 357 0.5517774 65 0.3939394 0.4303540
#    cancer
# tea     lower     upper      p.value
#   0        NA        NA           NA
#   b 0.4209577 1.0498032 8.281283e-02
#   g 0.2852035 0.6493768 9.265908e-05



ggplot(tea.df, aes(x=as.character(tea), 
                   y=freq, 
                   fill=as.character(cancer)))+
geom_col()+theme_classic()               

#4
chol<-data.frame(vacc=c(1,1,0,0), 
                 surv=c(1, 0, 1, 0), 
                 freq=c(55, 5, 15, 40))
summary(chol)
chol.tab<-xtabs(freq~vacc+surv, data=chol)
chisq.test(chol.tab)
# 	Pearson's Chi-squared test with Yates'
# 	continuity correction
# 
# data:  chol.tab
# X-squared = 47.289, df = 1, p-value =
# 6.124e-12

phi(chol.tab)
# [1] 0.66

# Here, be careful about the correlation (found) vs. 
# causality. The question was formulated in a causal way.
# Thus, it cannot be addressed by observational data. 
# An experiment would be needed.