R shiny not taking user input correctly

Question

Very basic R shiny issue Im sure. I want to run a linear regression on a toy dataset, use that to predict corresponding values based on used input, and plot a line to the regression line from the given and predicted values.

When I run this, I get the error: Listening on http://XXX Warning: 'newdata' had 1 row but variables found have 6 rows Error in segments(newData, 0, newData, newData.pred, col = "red") : invalid first argument

SO I realise it's how I'm handling the user input. I jsut want to take the numeric value they give, feed it to predict.lm and draw the resultant line. Id like text too if possible to give the predict.lm output.

ui.R

 library(shiny)
 (pageWithSidebar(
 headerPanel('Predicty'),
sidebarPanel
(
numericInput('clusters', 'Absorbance', 0.1, min = 0, max = 100, step = 0.00001)
),
mainPanel
(
  plotOutput('plot1')
)
))

server.R

 library(shiny)

##  Initial data and linear regression. 
conc <- c(.25, .5, 1, 1.5, 2, 2.5) 
abs <- c(.332, .631, .972, 1.201, 1.584, 1.912) 
results <- data.frame(abs,conc) 
abss <- lm(conc~abs, data = results)


shinyServer(function(input, output) {

output$plot1 <- renderPlot({


 newData.in <- reactive({
inFile <- input$clusters
inFile
 })

 newData <- data.frame(newData.in()) 
newData.pred <- predict.lm(abss, newData) 

plot(abs,conc) 
abline(abss)

segments(newData , 0, newData, newData.pred, col = "red")
 # segments(0, newData.pred[i], clusters()$centers, newData.pred[i], col = i)

  })

})

Answer 1

You don't need the reactive in your renderPlot , the function is already reactive .

predict.lm did not find any column names abs in your input which throws the first warning.

segments takes coordinates as arguments, you inputed a data.frame which explains the second error.

You could do:

 output$plot1 <- renderPlot({   
    newData <- data.frame(abs=input$clusters) 
    newData.pred <- predict.lm(abss, newData) 
    abline(abss)   
    segments(newData$abs , 0, newData$abs, newData.pred, col = "red")

  })

Also, you might want to change the xlim of your plot or the start value for your input because 0.1 is not on the plot