scale_x_binned() vs. POSIXct

[domq]

What I am trying to achieve:

Binned histograms rendered as Excel®-style grouped bar graphs, when the x axis is a timestamp.

This does exactly what I want when x is numeric:

ggplot(diamonds, aes(x = price, fill = cut)) + geom_bar(stat = "count", position = "dodge") + scale_x_binned()

The same approach doesn't work with POSIXct:

library(dplyr)
library(lubridate)
lakers %>% mutate(date = ymd(date)) %>%
ggplot(aes(x = date, fill = game_type)) + geom_bar(stat = "count", position = "dodge") + scale_x_binned()

Error: Binned scales only support continuous data

“Casting” the x into an integer is tantalizing close to what I want:

lakers %>% mutate(date = ymd(date)) %>%
ggplot(

    aes(x = int(date),  # Emphasis mine

        fill = game_type)) +
geom_bar(stat = "count", position = "dodge") + scale_x_binned()

... but of course, the labeling on the x axis could be better.

[domq]

Also,

lakers %>% mutate(date = as.Date(ymd(date))) %>%
    ggplot(aes(x = date, fill = game_type, width = 0.1)) +
    geom_bar(position = "dodge") +
    scale_x_binned(trans = "date")

Error: Invalid input: date_trans works with objects of class Date only