R plotly dropdown event for grouped boxplot - r

Question

I would like to use a plotly dropdown event to show different grouped boxplots, however I have not been able to achieve this as yet:
The first plot shows expected output for plotly with dropdown = "4" (obtained using ggplot). The second plot is what I get...
library(tidyverse)
library(plotly)
dat <- mtcars %>%
filter(cyl == 4 | carb == 4) %>%
group_by(cyl, carb, am) %>%
summarise(boxplot= list( setNames(boxplot.stats(disp)$stats,
c('lower_whisker','lower_hinge','median','upper_hinge','upper_whisker')) )) %>%
unnest_wider(boxplot) %>%
arrange(cyl, carb, am) %>%
ungroup() %>%
mutate_at(vars(cyl, carb, am), as.character)
cylinders <- unique(dat$cyl)
dat %>%
filter(cyl == 4) %>%
ggplot(aes(
x = carb,
lower = lower_hinge,
upper = upper_hinge,
middle = median,
ymin = lower_whisker,
ymax = upper_whisker,
colour = am)) +
geom_boxplot(stat = "identity")
p <- plot_ly(type = "box")
for(icyl in cylinders){
dataFilt <- filter(dat, cyl == icyl)
p <- add_trace(p,
visible = TRUE,
q1 = dataFilt$lower_hinge,
median = dataFilt$median,
q3 = dataFilt$upper_hinge,
lowerfence = dataFilt$lower_whisker,
upperfence = dataFilt$upper_whisker,
x = dataFilt$carb,
color = dataFilt$am,
name=icyl
)
}
p %>%
layout(boxmode = "group",
updatemenus = list(
list(
y = 0.8,
buttons = list(
list(label = cylinders[1],
method = "update",
args = list(list(visible = c(TRUE, FALSE, FALSE)))),
list(label = cylinders[2],
method = "update",
args = list(list(visible = c(FALSE, TRUE, FALSE)))),
list(label = cylinders[3],
method = "update",
args = list(list(visible = c(FALSE, FALSE, TRUE))))
))))

Answer 1

I wasn't able to get it to work with the loop, but I did get it to work as you had expected. Instead of using the variation where you did all the manual work to create the hinges, whiskers, etc., I used add_boxplot.
In layout you'll see that there are 6 T or F. That's because plotly doesn't save grouping as a function, it transforms the data into two separate traces. For example, the first two are cyl == "4" & am == "0" and cyl == "4" & am == "1"
cylinders <- unique(mtcars$cyl) # kept similar from original work
# added to simplify build
mtcars <- mtcars %>% mutate_at(vars(cyl, carb, am), as.character)
# made all cylinder options subplots
plot_ly() %>%
add_boxplot(x = ~carb, y = ~disp, color = ~am, colors = "Set2",
data = mtcars[mtcars$cyl == cylinders[1], ],
visible = T, inherit = F) %>% # visible
add_boxplot(x = ~carb, y = ~disp, color = ~am, colors = "Set2",
data = mtcars[mtcars$cyl == cylinders[2], ],
visible = F, inherit = F) %>% # invisible
add_boxplot(x = ~carb, y = ~disp, color = ~am, colors = "Set2",
data = mtcars[mtcars$cyl == cylinders[3], ],
visible = F, inherit = F) %>% # invisible
layout(updatemenus = list(
list(
y = 0.8,
buttons = list(
list(label = cylinders[1], # four cyl
method = "restyle",
args = list("visible", list(T, T, F, F, F, F))),
list(label = cylinders[2], # six cyl
method = "restyle",
args = list("visible", list(F, F, T, T, F, F))),
list(label = cylinders[3], # eight cyl
method = "restyle",
args = list("visible", list(F, F, F, F, T, T)))
) # end buttons
)) # end updatedmenus list list
) # end layout
Here's the 4 cyl plotly and your original 4 cyl ggplot