plot suggestions for two categorical vars

[bragks]

I'm looking for suggestions/advice on how to plot two categorical (but somewhat numeric...) values. The predictor, x, is a 5 point Likert type scale where score 1 is just 1-2 binned together and y is the outcome of a test performed (not shown).

It feels like the result is more obvious in Plot 2, but I'm not sure if using a scatterplot for this type of data is "acceptable"? Does anyone have a suggestion on how I could plot this instead of a barchart?

I had a look at https://www.data-to-viz.com, but I'm none the wizer...

library(tidyverse)

Df <- tibble::tribble(
  ~x, ~y,
  4,  2,
  1,  1,
  3,  2,
  1,  0,
  1,  0,
  3,  0,
  5,  2,
  5,  4,
  3,  2,
  3,  0,
  5,  2,
  4,  3,
  1,  0,
  4,  1,
  1,  0,
  1,  0,
  4,  1,
  3,  0,
  5,  3,
  1,  0,
  1,  0,
  1,  0,
  1,  0,
  4,  0,
  1,  0,
  1,  0,
  5,  5,
  1,  0,
  5,  0,
  3,  3,
  1,  2,
  4,  2,
  1,  0,
  5,  5,
  4,  4,
  5,  4,
  3,  0,
  1,  0,
  1,  0,
  4,  4,
  1,  0,
  1,  0,
  5,  4,
  1,  0,
  1,  1,
  5,  3,
  5,  4,
  5,  5,
  1,  4,
  5,  3,
  1,  0,
  1,  0,
  5,  1,
  1,  0,
  1,  0,
  4,  4,
  1,  0,
  1,  0,
  1,  0,
  5,  3,
  5,  2,
  1,  0,
  1,  0,
  1,  1,
  5,  3,
  1,  0,
  5,  4,
  1,  2,
  3,  1,
  1,  0,
  1,  0,
  1,  0,
  5,  3,
  1,  0,
  1,  0,
  5,  5,
  1,  1,
  1,  0,
  5,  3,
  1,  0,
  5,  2,
  1,  0,
  4,  1,
  5,  3,
  3,  0,
  3,  0,
  1,  0,
  5,  2,
  1,  3,
  5,  2,
  5,  5,
  5,  3,
  3,  2,
  1,  0,
  4,  2,
  5,  3,
  5,  4,
  3,  0,
  1,  0,
  5,  0
) %>% 
  mutate_if(is.numeric, factor)

#Plot 1
Df %>% 
  ggplot(aes(x)) +
  geom_bar(aes(fill = y), position = "dodge") +
  scale_y_log10()

#Plot 2
Df %>% 
  ggplot(aes(x, y)) +
  geom_jitter(aes(color = y), height = 0.15, width = 0.15, size = 4, alpha = 0.5)

^{Created on 2019-03-04 by the reprex package (v0.2.1)}

[cpsievert]

The answer depends on what question (i.e., comparison) you'd like to ask of the data. In this case, I suspect you likely want to start by comparing outcomes conditional on x. In other words, you might want to use position = "fill" to compare the relative rather absolute frequencies

Df %>% 
  ggplot(aes(x)) +
  geom_bar(aes(fill = y), position = "fill")

The downside to that graphic is that you no longer have a sense of how many outcomes fall under each x category. One way to add that information is to scale the width of each x category proportional to the number of observations in each category. The mosaic plot is a general visualization framework that would allow you to do this. There are several R packages that implement mosaic plots -- I'm biased, but I like ggmosaic since I can query the frequency that each rectangle encodes via plotly

library(ggmosaic)
ggplot(Df) + geom_mosaic(aes(x = product(x), fill = y))
plotly::ggplotly()

In this particular case, I think the mosaic plot is useful, but in general, they can become difficult to interpret fairly quickly, so depending on your audience, you may want to avoid them. Here is a list of relevant reading if you're interested in learning more:

https://homepage.divms.uiowa.edu/~luke/classes/STAT4580/morecat.html

Mosaic plot

prodplots.pdf

are_mosaic_plots_worthwhile.pdf