Tooltip removes regression line ggplotly

Question

I have the following code in server.R:

library(shiny)
library(plotly)

art.data <- read.csv("data1.csv", stringsAsFactors = FALSE)

shinyServer(function(input, output) {

output$distPlot <- renderPlotly({
    col.str <- paste0(input$colspa, ".", input$rgbchoice, ".median")
    p <- ggplot(art.data, aes(x = year, y = art.data[[col.str]], text = paste0(artist, "<br>", art))) + geom_point(size = 1) + xlab("Year") + stat_smooth(method = loess, se = FALSE)
    ggplotly(p , tooltip = "text")
})
})

If I remove tooltip then the regression line is there in output graph but with tooltip included, regression line does not appear in the graph. Any solutions to have both the things together?

With tooltip

Without tooltip in ggplotly() and text in aes()

Answer 1

The aesthetic text needs the aesthetic group.
Let start considering this example:

art.data <- data.frame(year=1430:1490, y=rnorm(61), 
            artist=sample(LETTERS,61, replace=T), art=sample(letters,61, replace=T))
col.str <- 2

library(ggplot2)
ggplot(art.data, aes(x = year, y = art.data[[col.str]], 
                     text=paste0(artist, "<br>", art))) + 
    geom_point(size = 1) + 
    stat_smooth(method = loess, se = FALSE) +
    xlab("Year") 

In this case the loess line is not plotted on the graph:

Adding group=1 in the ggplot aesthetics solves the problem:

p <- ggplot(art.data, aes(x = year, y = art.data[[col.str]], 
                     text=paste0(artist, "<br>", art), group=1)) + 
    geom_point(size = 1) + 
    stat_smooth(method = loess, se = FALSE) +
    xlab("Year") 
p

and ggplotly now works nicely:

library(plotly)
ggplotly(p, tooltip="text")