Correlation demonstration

"Andrea Harnos"
18.08.2020

Goal of the shiny app

In teaching correlation in class, to show, how the scatterplot of points generated from bivariate normal distribution look like.
Number of data points and
the value of Pearson correlation can be given.
Scatterplot with calculated correlation is shown.

ui.r

library(shiny)
shinyUI(fluidPage(
  titlePanel("Demonstration of correlation"),
  sidebarLayout(
    sidebarPanel(
      sliderInput("numeric", "How Many Random Numbers Should be generated?", 
                   value = 100, min = 1, max = 1000, step = 1),
      sliderInput("corr", "Set correlation",                            -1, 1, value = 0.5, step=0.1),
      h4("The application generates random numbers from bivariate normal distribution (mean = 0, variance = 1). The number of points and  the value of correlation can be adjusted. Scatterplot of the generated random numbers are shown.")
  ),
  mainPanel(
  plotOutput("plot1")))))

The scatterplot

server.R

library(shiny)
library(MASS)
shinyServer(function(input, output) {
renderPlot({
  number_of_points <-input$numeric
  inpcorr <-input$corr
  mu <- rep(0,2)
  Sigma <- rbind(c(1,inpcorr),c(inpcorr,1))
  rawvars <- mvrnorm(n=number_of_points, mu=mu, Sigma=Sigma)
  plot(rawvars[,1],  rawvars[,2], xlab="X",ylab="Y",
       xlim = c(-4, 4), ylim = c(-4, 4), las=1)
  text(-3,3.5,paste("Number of points =",number_of_points))
  text(2,3.5,paste("Correlation given =",inpcorr))
  calc_cor<-round(cor( rawvars[,1], rawvars[,2]),2)
  text(2,3,paste("Calculated correlation =",calc_cor))
})