2 - Wilcoxon with Boxes.Rmd

---
title: "Bits of Code"
author: "Amir Nakar"
date: "10/1/2020"
output: word_document
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
```

```{r, include=FALSE}
load(file = "~/MorK.2020/REnviron.RData")
#### Load the data ####
library(openxlsx)
library(reshape2)
library(ggplot2)
library(Hmisc) 
library(dplyr)
library(car)
library(caret)
library(readxl)
library(ggsignif)
library(tidyr)
library(vegan)
library(purrr)
library(egg)

```
 

```{r message=FALSE}
# Significance testing: 
# We will use the wilcoxon Ranked Sum Test as a non-parametric test:
colnames(data)

#### Function explained ####
# This function will work as follows: 
# Input: two variable names from your data set
# Output: Boxplot with Wilcoxon Rank Sum Test statistic printed on it


# Notes: 
# 1. You can ignore the warnings: 
  # No id variables; using all as measure variables
  # Removed 2 rows containing non-finite values
# It's just the way it handles NA lines. Which you have

# 2. Box plots: 
# The middle line is the median
# 
# The lower and upper hinges correspond to the first and third quartiles
#  
# The whiskers extend from the hinge to the furthest value but no further than 1.5 * IQR from the hinge 
# (where IQR is the inter-quartile range, or distance between the first and third quartiles). 

# Outliers: Points beyond the whisker limit


                              wilcox.plus.boxes <- function(var1, var2) {  # Defines the function requiring two parameters
data.for.function = data[,c(var1,var2)]      # creates a dataframe from your data which is only the two parameters you chose
colnames(data.for.function) = c(var1,var2)   # Gives the relevant columns names (for later)

varname = gsub("\\..*","",var1)
levs    = c("New","Old")

result = wilcox.test(                        # Calculates the wilcoxon test between the two variables
  data.for.function[,1],
  data.for.function[,2]
  ) 

mean.diff = mean(data.for.function[,1], na.rm = T)-mean(data.for.function[,2],na.rm = T)

melted = melt(data.for.function)             # Transforms the data for ggplot

plot = ggplot(                               # This is a very long function to create the plot
  melted, aes(x=variable,y=value,fill=variable)) + # Defines the releant data
  geom_boxplot()+                                 # Defines the plot type as Box Plot
  theme(panel.background = element_rect(fill="white")) + # Makes it visually clean
  labs(x = varname,y="Time [Mins]") +                  # Adds X, Y axis labels
  theme(panel.grid.major = element_line(colour = "gray")) + # adds gray lines on the Y axis
  theme(panel.grid.minor = element_line(colour = "gray")) + # does the same
  guides(fill=FALSE) +                                      # Removes the legend (not needed for this)
  labs(                                                     # Startes to create the complicated titles
    title = paste("Comparison of", varname, sep = " "), # Main title: Comparison between the two variables
    
    
    
    
    subtitle = paste(                                                # Subtitle: gives the different statistics pasted together
      "Result of the Wilcoxon Rank Sum Test: \ntest statistic = ", 
      result$statistic, "\n p-value = ", 
      formatC(result$p.value, format = "g", digits = 2), 
      "\n mean difference = ", round(mean.diff, digits = 1), "minutes", sep = " " )
      ) + 
  scale_x_discrete(breaks = c(var1,var2), labels=c("New","Old")) +
  
  geom_signif(comparisons = list(c(var1, var2)),             # Creates a significance star when the pvalue<0.05
  test="wilcox.test", map_signif_level =  c("*"=0.05))
  

plot                                                         # Saves the final plot 

  }

names = colnames(data) # This is for my own reference to know which variables need to be tabulated

# Calculations:
# Each Q is a question from your list
Q1 = wilcox.plus.boxes(names[33],names[32]) #SUM
Q2 = wilcox.plus.boxes(names[19],names[18]) # Appointments1
Q3 = wilcox.plus.boxes(names[27],names[26]) # Mirshm
Q4 = wilcox.plus.boxes(names[29],names[28]) # Payments
Q5 = wilcox.plus.boxes(names[31],names[30]) # Update
Q6 = wilcox.plus.boxes(names[21],names[20]) # Appointments 2
Q7 = wilcox.plus.boxes(names[25],names[24]) # Medical File
Q8 = wilcox.plus.boxes(names[23],names[22]) # Blood Test

```

```{r message=FALSE}
Q1
Q2
Q3
Q4
Q5
Q6
Q7
Q8

FIG3 = ggarrange(Q1,Q2,Q3,Q4,Q5,Q6,Q7,Q8, ncol = 4)
```


# Calculating Wilcoxon either one or two way for report: 
``` {r}
colnames(data)

wilcox.test(                        # Calculates the wilcoxon test between the two variables
  data[,22],
  data[,23], alternative = "less")
```


# And now for all of them: 


```{r message=FALSE}
# Significance testing: 
# We will use the wilcoxon Ranked Sum Test as a non-parametric test:
colnames(data)

#### Function explained ####
# This function will work as follows: 
# Input: two variable names from your data set
# Output: Boxplot with Wilcoxon Rank Sum Test statistic printed on it


# Notes: 
# 1. You can ignore the warnings: 
  # No id variables; using all as measure variables
  # Removed 2 rows containing non-finite values
# It's just the way it handles NA lines. Which you have

# 2. Box plots: 
# The middle line is the median
# 
# The lower and upper hinges correspond to the first and third quartiles
#  
# The whiskers extend from the hinge to the furthest value but no further than 1.5 * IQR from the hinge 
# (where IQR is the inter-quartile range, or distance between the first and third quartiles). 

# Outliers: Points beyond the whisker limit


                              wilcox.plus.boxes <- function(var1, var2) {  # Defines the function requiring two parameters
data.for.function = data[,c(var1,var2)]      # creates a dataframe from your data which is only the two parameters you chose
colnames(data.for.function) = c(var1,var2)   # Gives the relevant columns names (for later)

varname = gsub("\\..*","",var1)
levs    = c("New","Old")

result = wilcox.test(                        # Calculates the wilcoxon test between the two variables
  data.for.function[,1],
  data.for.function[,2], alternative = "less"
  ) 

mean.diff = mean(data.for.function[,1], na.rm = T)-mean(data.for.function[,2],na.rm = T)

melted = melt(data.for.function)             # Transforms the data for ggplot

plot = ggplot(                               # This is a very long function to create the plot
  melted, aes(x=variable,y=value,fill=variable)) + # Defines the releant data
  geom_boxplot()+                                 # Defines the plot type as Box Plot
  theme(panel.background = element_rect(fill="white")) + # Makes it visually clean
  labs(x = varname,y="Time [Mins]") +                  # Adds X, Y axis labels
  theme(panel.grid.major = element_line(colour = "gray")) + # adds gray lines on the Y axis
  theme(panel.grid.minor = element_line(colour = "gray")) + # does the same
  guides(fill=FALSE) +                                      # Removes the legend (not needed for this)
  labs(                                                     # Startes to create the complicated titles
    title = paste("Comparison of", varname, sep = " "), # Main title: Comparison between the two variables
    
    
    
    
    subtitle = paste(                                                # Subtitle: gives the different statistics pasted together
      "Result of the Wilcoxon Rank Sum Test: \ntest statistic = ", 
      result$statistic, "\n p-value = ", 
      formatC(result$p.value, format = "g", digits = 2), 
      "\n mean difference = ", round(mean.diff, digits = 1), "minutes", sep = " " )
      ) + 
  scale_x_discrete(breaks = c(var1,var2), labels=c("New","Old")) +
  
  geom_signif(comparisons = list(c(var1, var2)),             # Creates a significance star when the pvalue<0.05
  test="wilcox.test", map_signif_level =  c("*"=0.05))
  

plot                                                         # Saves the final plot 

  }

names = colnames(data) # This is for my own reference to know which variables need to be tabulated

# Calculations:
# Each Q is a question from your list
Q1alt = wilcox.plus.boxes(names[33],names[32]) #SUM
Q2alt = wilcox.plus.boxes(names[19],names[18]) # Appointments1
Q3alt = wilcox.plus.boxes(names[27],names[26]) # Mirshm
Q4alt = wilcox.plus.boxes(names[29],names[28]) # Payments
Q5alt = wilcox.plus.boxes(names[31],names[30]) # Update
Q6alt = wilcox.plus.boxes(names[21],names[20]) # Appointments 2
Q7alt = wilcox.plus.boxes(names[25],names[24]) # Medical File
Q8alt = wilcox.plus.boxes(names[23],names[22]) # Blood Test

```

```{r message=FALSE}
Q1
Q2
Q3
Q4
Q5
Q6
Q7
Q8

FIG3 = ggarrange(Q1alt,Q2alt,Q3alt,Q4alt,Q5alt,Q6alt,Q7alt,Q8alt, ncol = 4)
```