#3 Changed the data from iris to the tki demo

vignettes/03_analysis_modelling.Rmd

@@ -24,6 +24,7 @@ library(ggplot2)
 library(gridExtra)
 library(grid)
 library(tidyverse)
+library(ggpubr)
 data(iris)
 
 ```
@@ -32,11 +33,11 @@ data(iris)
 
 ## What we cover
 
->- Linear Regression
->- Multiple Linear Regression 
->- Logistic Regression
+- Linear Regression
+- Multiple Linear Regression 
+- Logistic Regression
 
-```{r echo=FALSE, error=FALSE, message=FALSE, warning=FALSE, 'class="centre", out.extra=, style="width:, warnings=FALSE}
+```{r echo=FALSE, error=FALSE, message=FALSE, warning=FALSE, out.extra = 'class="centre" style="width: 500px;"', warnings=FALSE}
 setwd("/Users/srauschert/Desktop/Work/20.) Git_GitHub/RWorkshop/")
 tki_demo <- read_csv("data/demo.csv")
 
@@ -58,9 +59,9 @@ ggplot( aes(day2, day3)) +
 
 In a linear regression, we aim to find a model: <br /> 
 
->- that represents our data and 
+- that represents our data and 
 
->- can give information about the association between our variables of interest.
+- can give information about the association between our variables of interest.
 
 The command in R for a linear model is <br />
 
@@ -73,58 +74,72 @@ The Iris data set consists of information about three different species of iris
 
 It holds information on:
 
->- Sepal length
+- Sepal length
 
->- Sepal width
+- Sepal width
 
->- Petal length
+- Petal length
 
->- Petal width
+- Petal width
 
 ## Data set summary
 Let's first have a look at the summary table of the Iris data set, by using the <code>summary()</code> command:
 
-```{r, echo = FALSE, results='asis',out.extra = 'class="centre" style="width: 500px;"'}
-kable(summary(iris[,c(1:4)]))
+```{r, echo = FALSE, results='asis',out.extra = 'class="centre" style="width: 100px;"',warning=FALSE}
+kable(summary(tki_demo[,c(6:8)]))
 ```
 
 #Visualisation of data distributions
 ##Helpful plots before modelling
 Before we start with the linear regression model, we need to get an idea of the underlying data and its distribution.
 We know that the linear regression has the assumtptions:
 
->-
+-
 
 
 ## QQ-plot: 
-```{r, echo=FALSE, out.extra = 'class="centre" style="width: 700px;"'}
+```{r, echo=FALSE, out.extra = 'class="centre" style="width: 700px;"', warning=FALSE}
 library(tidyr)
-data(iris)
-iris_long <- gather(iris, Specification, measurement, Sepal.Length:Petal.Width, factor_key=TRUE)
 
-ggplot(iris_long, aes(sample=measurement, color=Specification))+stat_qq()
+tki_demo %>%
+  filter(day2 < 100) %>%
+  gather(Days, measurement, day1:day3, factor_key=TRUE) %>%
+  ggplot( aes(sample=measurement, color=Days))+stat_qq()
 
 ```
 
 ## Boxplots to check for outliers
 
 
-```{r echo = FALSE, out.extra = 'class="centre" style="width: 700px;"'}
+```{r echo = FALSE, out.extra = 'class="centre" style="width: 700px;"',warning=FALSE}
 
-  
+with_out <- tki_demo %>%
+  #filter(day2 < 100) %>%
+  gather(Days, measurement, day1:day3, factor_key=TRUE) %>%
+  ggplot(aes(y=measurement,x=Days, fill=Days)) +
+  labs(title = "Days: 1 to 3 with outlier", x = "", y = "Measurment") +
+  geom_boxplot() +
+  scale_color_telethonkids("light") +
+  theme_minimal()
 
-ggplot(iris_long, aes(y=measurement,x=Specification, col=Specification)) +
-  labs(title = "Iris Specifications", x = "", y = "Measurment in cm") +
+no_out <- tki_demo %>%
+  filter(day2 < 100) %>%
+  gather(Days, measurement, day1:day3, factor_key=TRUE) %>%
+  ggplot(aes(y=measurement,x=Days, fill=Days)) +
+  labs(title = "Days: 1 to 3 outlier removed", x = "", y = "Measurment") +
   geom_boxplot() +
   scale_color_telethonkids("light") +
   theme_minimal()
 
+
+ggarrange(with_out, no_out, ncol=2, common.legend = TRUE, legend=FALSE )
+
 ```
 
 
 ## Plot the variables
 
-```{r, echo = FALSE, out.extra = 'class="centre" style="width: 700px;"'}
+```{r, echo = FALSE, out.extra = 'class="centre" style="width: 700px;"',warning=FALSE}
 data(iris)
 plot1 <- ggplot(iris, aes(Petal.Width, Petal.Length)) +
   labs(title = "Petal", x = "Petal Width", y = "Petal Length") +
@@ -164,11 +179,11 @@ Let's now perform a linear regression model in R.
 
 <code>lm(Petal.Length~Petal.Width, data=iris)</code>
 
->- As said before, the first argument in the code is **<em>y</em>**, our outcome variable or <em>dependent variable</em>. In this case it is **<em>Petal.Length</em>**.
+- As said before, the first argument in the code is **<em>y</em>**, our outcome variable or <em>dependent variable</em>. In this case it is **<em>Petal.Length</em>**.
 
->- The second Argument is **<em>x</em>**, the <em>independent variable</em>. In our case: **<em>Petal.Width</em>**.
+- The second Argument is **<em>x</em>**, the <em>independent variable</em>. In our case: **<em>Petal.Width</em>**.
 
->- We also specify the data set that holds the variables we specified as **<em>x</em>** and **<em>y</em>**.
+- We also specify the data set that holds the variables we specified as **<em>x</em>** and **<em>y</em>**.
 
 ##Linear Regression Results
 Now we want to look at the results of the linear regression. So how do we get the <em>p-value</em> and <em>\(\beta\)-coefficient</em> for the association?