27/05/2019
This Shiny app is a predictor of mpg based on cylinder numbers(cyl) and horse power (hp) from mtcars dataset. It is needed to enter parameters on the left side, and on the right side there will be a prediction, and a plot of a linear regression. The project consists of two files: ui.R and server.R
ui.R file consists of a sidebar part and main part. The side part consists of two parts, - a numeric input part for the number of hp with a step of 1 (defaulted to 145 due to mean being around 146) - and an optional cyl part (by default unspecified, drop down list due to only 3 possible different values).
On the right are the results.
mean(mtcars$hp)
## [1] 146.6875
unique(mtcars$cyl)
## [1] 6 4 8
library(shiny)
shinyUI(fluidPage(
titlePanel("Mpg prediction from mtcars dataset"),
sidebarLayout(
sidebarPanel(
h3(helpText("Choose values:")),
numericInput("hp", label = h4("hp"), step = 1, value = 145),
selectInput("cyl", label = h4("cyl"),
choices = list("unspecified" = "*", "4" = "4", "6" = "6", "8" = "8"))
),
mainPanel(
h4("Predicted value of this diamond is:"),
h3(textOutput("result")),
h4(""),
h4("Regression results:"),
plotOutput("distPlot")
)
)
The server part consists of two parts, the first part being the plot, and the second being the prediction. The plot has been done with ggplot2 (smoothed). The prediction part can be run both with and without cylinder number specified.
library(shiny)
library(ggplot2)
library(dplyr)
cars <- mtcars
shinyServer(function(input, output) {
output$distPlot <- renderPlot({
# select diamonds depending of user input
cars <- filter(mtcars, grepl(input$cyl, cyl))
# build linear regression model
fit <- lm( mpg~hp, cars)
# predicts the price
pred <- predict(fit, newdata = data.frame(hp = input$hp, cyl = input$cyl))
# Drow the plot using ggplot2
plot <- ggplot(data=cars, aes(x=hp, y = mpg))+
geom_point(aes(color = cyl), alpha = 0.3)+
geom_smooth(method = "lm")+
geom_vline(xintercept = input$hp, color = "green")+
geom_hline(yintercept = pred, color = "red")
plot
})
output$result <- renderText({
# renders the text for the prediction below the graph
cars <- filter(cars, grepl(input$cyl, cyl))
fit <- lm( mpg~hp, cars)
pred <- predict(fit, newdata = data.frame(hp = input$hp, cyl = input$cyl))
res <- paste(round(pred, digits = 2), "$")
res
})
})