Differences in behaviour between the ggplot and tidier implementations of geom_bar() and geom_col().

[JasonPekos]

Looking at the R bar plot documentation, there are significant differences between the Julia and R bar plot implementations. Some of these — related to sideways bar plots — are pretty significant.

---

Consider the following ggplot (R) code:

ggplot(mpg) + geom_bar(aes(y = class))

The corresponding Julia code seems like it should be:

using RDatasets
using TidierPlots

cars = dataset("datasets", "mtcars")

ggplot(cars) + 
    geom_bar(@aes(y = Model))

But Tidier fails here, requiring an x aesthetic.

geom_col() is a little weirder. In R:

ggplot(mpg, aes(y = class, x = cyl)) + geom_col()

However in TidierPlots:

ggplot(cars) + 
    geom_col(@aes(y = Model, x = Cyl))

I think at least the geom_col behaviour is concerning.

[rdboyes]

Finally making progress on this issue - in 0.6 (not released yet):

using TidierPlots
using DataFrames

df = DataFrame(cat = ["left", "left", "left",
 "middle", "middle", "middle", "right", "right", "right"],
       height = 0.1:0.1:0.9,
       grp = [1, 2, 3, 1, 2, 3, 1, 2, 3],
       grp1 = [1, 2, 2, 1, 1, 2, 1, 1, 2],
       grp2 = [1, 1, 2, 1, 2, 1, 1, 2, 1]
)

ggplot(df, yticks = (1:3, ["bottom", "middle", "top"])) + 
    geom_col(@aes(cat, height, color = grp, bar_labels = height), 
        position = "dodge", direction = "x") + labs(title = "Dodged Bars") + 
    theme_dark()

[rdboyes]

The geom_bar behaviour is fixed now, but I think the geom_col behaviour in R is actually ... bad? It's not actually clear to me what the plot in the example should look like in order to make sense. In any case:

using TidierPlots
using DataFrames
using CSV
    
cars = CSV.read("scratch/auto-mpg.csv", DataFrame)
    
ggplot(cars) + 
    geom_bar(@aes(y = class))

Produces (in 0.6.0):