geom_point_interactive issue when passed alongside geom_boxplot

Question

I wonder if you can help me.

I often use the splendid ggiraph, but have recently encountered an issue where it doesn't seem to work when one of its interactive geoms is paired with another geom (either from ggiraph or ggplot2). Let me explain with my example below and start with the data.

library(ggiraph)
library(tidyverse)
dput(data)

structure(list(moniker = c("OU", "IN", "AI", "HR", "T ", "LA", 
"AJ", "WC", "BE", "NV"), monthly = c(-0.03489, -0.028034, -0.0374, 
-0.0479, -0.032485, -0.0332, -0.045954, -0.048032, -0.0446, -0.027724
), quarterly = c(-0.096452525515629, -0.0642786819909756, -0.12256466734, 
-0.112914230016, -0.121439751549064, -0.12779609, -0.103490246054841, 
-0.152770470354658, -0.12167068725, -0.0993823354535133), yearly = c(-0.107231010977261, 
-0.00722368758143395, -0.088470240158463, -0.0817767947388421, 
-0.136265993838053, -0.0970216743424386, -0.092159717986092, 
-0.174532024103611, -0.131514750319878, -0.0866996056676958)), class = c("tbl_df", 
"tbl", "data.frame"), row.names = c(NA, -10L))

In other words, the data looks like this:

glimpse(data)

Observations: 10
Variables: 4
$ moniker   <chr> "OU", "IN", "AI", "HR", "T ", "LA", "AJ", "WC", "BE", "NV"
$ monthly   <dbl> -0.034890, -0.028034, -0.037400, -0.047900, -0.032485, -0...
$ quarterly <dbl> -0.09645253, -0.06427868, -0.12256467, -0.11291423, -0.12...
$ yearly    <dbl> -0.107231011, -0.007223688, -0.088470240, -0.081776795, -...

When I do things conventionally, all works fine:

data %>% 
  gather(key = "timeframe", value = "return", -moniker) %>% 
  mutate(timeframe = as_factor(timeframe)) %>% 
  ggplot(
    aes(
      x = timeframe, 
      y = return
      )
    ) + 
  geom_hline(yintercept = 0) + 
  geom_point(position = position_jitter(w = 0.15, h = 0)) + 
  geom_boxplot()

... generates:

However, when I try to use geom_point_interactive instead of geom_point:

data %>% 
  gather(key = "timeframe", value = "return", -moniker) %>% 
  mutate(timeframe = as_factor(timeframe)) %>% 
  ggplot(
    aes(
      x = timeframe, 
      y = return,
      tooltip = moniker
      )
    ) + 
  geom_hline(yintercept = 0) + 
  geom_point_interactive(position = position_jitter(w = 0.15, h = 0)) + 
  geom_boxplot()

... I get the following:

(Note that I couldn't upload an svg file here, but can confirm that the tooltips from the geom_point_interactive work fine.) Furthermore, the same behaviour seems to happen if I replace geom_boxplot with geom_boxplot_interactive.

So ... to get to my question: Do you know of a way to preserve the tooltips on the points, whilst having a boxplot that looks like it does in the first picture above?

As always, many thanks for looking.

score 2 · Accepted Answer · answered Feb 13 '19 at 13:17

The problem is that the geom_boxplot is inheritng tooltip = moniker which means each moniker will get its own box, to avoid this move the tooltip aesthetic to geom_point_interactive like so:

my_data %>% 
  gather(key = "timeframe", value = "return", -moniker) %>% 
  mutate(timeframe = as_factor(timeframe)) %>% 
  ggplot(
    aes(
      x = timeframe, 
      y = return
    )
  ) + 
  geom_hline(yintercept = 0) + 
  geom_boxplot() + 
  geom_point_interactive(aes(tooltip = moniker), 
                         position = position_jitter(w = 0.15, h = 0))

geom_point_interactive issue when passed alongside geom_boxplot

1 Answers1