CODEWITHSUNDEEP
rggplot2aeslegendfactors
WildParamecium
WildParamecium

Reputation: 3

ggplot2 is merging two different legend labels in a combined one

I'm trying to produce a scatter plot with ggplot for a shiny web app, to show the relationship between two variables(age and platelets, both numeric). Moreover, I want to add two additional dimensions to show possible differences in the linear relationship, plotted with color and shape (diabetes and anemia, both factor variables with possible values 0 or 1).

#selected() is a reactive function that apply filters to the dataframe
ggplot(data = selected(), aes(x=age, y=platelets, shape=diabetes, color=anaemia)) + geom_point()

This chunk of code is returning the following legend:

diabetes
0
(0,1)
(1,0)
1

And I was expecting the following output instead:

diabetes
0
1

anaemia
0
1

I have been reading at ggplot documentation but I didnt find a way to solve this problem. I really appreciate possible help to solve the problem.

Edit: Dataset: https://www.kaggle.com/andrewmvd/heart-failure-clinical-data

selected() function body

selected <- reactive({
    heart %>% 
      select(c(input$variables, response))
      filter(age >= input$age[1], age <= input$age[2])
  })

Upvotes: 0

Views: 240

Answers (1)

Brenton M. Wiernik
Brenton M. Wiernik

Reputation: 1316

ggplot2's default behavior is to have separate guides for each variable like you want. See the example below with code that is equivalent to what you have written.

library(ggplot2)

ggplot(data = mtcars) +
  aes(x = wt, y = mpg, shape = factor(am), color = factor(vs)) + 
  geom_point()

Created on 2021-05-30 by the reprex package (v2.0.0)

Is the code you showed actually what you are running that produces a different legend? If so, please post a reproducible example showing what happens.

Upvotes: 1

Related Questions