Search code examples
rggplot2r-plotlyggplotly

Plotly hovering on tiles in the background of a cloud of points - how to show the tooltip?

I'm converting a ggplot2 plot to plotly. The plot consists of a tile layer (in the background) and a point layer (in the foreground). I would like to have tooltips when hovering on the tiles.

The code below mostly gets me what I am looking for. When I hover on tiles in "points free" zones, the desired tooltip appears. However, when I hover in areas with a high density of dots, the tooltips do not appear.

I thought that playing with the layerData parameter in the ggplotly call might help, but that was not the case.

library(ggplot2)
library(dplyr)
library(plotly)

set.seed(1)
dat_points <- data.frame(x = rnorm(100), y = rnorm(100))
dat_tiles <- expand.grid(tx = -3:3, ty = -3:3)
dat_tiles$val <- rnorm(nrow(dat_tiles))
dat_tiles$label <- sample(LETTERS[1:5], nrow(dat_tiles), replace = T)

p <- ggplot() +
  geom_tile(data = dat_tiles, aes(x = tx, y = ty, fill = val, text = label)) +
  geom_point(data = dat_points, aes(x = x, y = y), alpha = .5)

gg <- ggplotly(p, tooltip = "text")
gg

I would like that hovering on high density areas (e.g. 0, 0) would bring up tooltips with the same promptness as in low density areas.

EDIT: added static image of the plot.

static version of the plot

Solution

  • You can switch the order of you layers in p and because of how ggplotly() constructs from a ggplot object, you get an identical looking plot, but with the desired tooltip behavior!

    p <- ggplot() +
  geom_point(data = dat_points, aes(x = x, y = y), alpha = 1) +
  geom_tile(data = dat_tiles, aes(x = tx, y = ty, fill = val, text = label))
p # this looks bad

gg <- ggplotly(p, tooltip = "text")
gg # but this looks good!