Playing with mtcars data

Mohammad Ullah
01/29/2018

Summary

In this presentation, I am going to explore the functionality of the shiny app. The app does some data analysis upon users command on mtcars data set. It contains two tabpanels.

In this first panel, it asks the user to choose three variables to make a plot with a linear regression line.
In the second panel, a comparison between two prediction model (rpart and random forest) is implemented.

app link: https://moeensaiket.shinyapps.io/Playwithmtcarsdata/ code link: https://github.com/mohammadullah/Shiny-App-/

First Panel plotting Code on server

cars1 <- mtcars
cars1$cyl <- as.factor(cars1$cyl)
cars1$vs <- as.factor(cars1$vs)
cars1$am <- as.factor(cars1$am)
cars1$gear <- as.factor(cars1$gear)
selecteddata <- reactive({
  cars1[, c(input$var1, input$var2, input$var3)]
  }) 
output$plot1 <- renderPlot({
  df1 <- selecteddata()
  p <- ggplot(df1, aes(x = df1[,1], y = df1[,2], color = df1[,3]))
  p <- p + geom_point(size = 5) + geom_smooth(method = "lm")
  p <- p + labs(x = names(df1)[1], y = names(df1)[2], color = names(df1)[3])
  p
  })

First default Panel plot (hp vs mpg)

library(ggplot2)
cars1 <- mtcars[, c("mpg", "hp", "cyl")]
cars1$cyl <- as.factor(cars1$cyl)

p <- ggplot(cars1, aes(x = cars1[,1], y = cars1[,2], 
                       color = cars1[,3]))
p <- p + geom_point(size = 5) + geom_smooth(method = "lm")
p <- p + labs(x = names(cars1)[1], y = names(cars1)[2], 
              color = names(cars1)[3])

p

plot of chunk unnamed-chunk-3

Comparison between rpart and random forest model

Use

Select variable = “cyl”
Select predictor variable = “wt”, “gear”, “mpg”
check both models

alt text