rggplot2scatterpie

Scatterpie doesn't recognize column name from data

I want to plot 8 data points using scatterpie. However, the function is throwing an error that it can't find the object corresponding to the column with values for y-coordinates.

My Data 

library(tidyverse)
library(scatterpie)

my_df <- structure(list(day_in_july = 13:20, yes_and_yes = c(0.611814345991561, 
0.574750830564784, 0.593323216995448, 0.610539845758355, 0.650602409638554, 
0.57429718875502, 0.575971731448763, 0.545454545454545), yes_but_no = c(0.388185654008439, 
0.425249169435216, 0.406676783004552, 0.389460154241645, 0.349397590361446, 
0.42570281124498, 0.424028268551237, 0.454545454545455)), row.names = c(NA, 
-8L), class = c("tbl_df", "tbl", "data.frame"))  

> my_df

## # A tibble: 8 x 3
## day_in_july yes_and_yes yes_but_no
##         <int>       <dbl>      <dbl>
## 1          13       0.612      0.388
## 2          14       0.575      0.425
## 3          15       0.593      0.407
## 4          16       0.611      0.389
## 5          17       0.651      0.349
## 6          18       0.574      0.426
## 7          19       0.576      0.424
## 8          20       0.545      0.455

Plotting the data using scatterpie fails

I followed the code from the documentation, but it still isn't working for me.

ggplot() + 
  geom_scatterpie(aes(x = day_in_july, y = yes_but_no), 
                           data = my_df, 
                           cols = colnames(my_df)[2:3])  


## Error in FUN(X[[i]], ...) : object 'yes_but_no' not found

I've tried converting from tibble to a data.frame beforehand, but didn't solve the problem.

By the way, setting y as a constant (e.g., 2) works:

ggplot() + 
  geom_scatterpie(aes(x = day_in_july, y = 2), 
                           data = my_df, 
                           cols = colnames(my_df)[2:3])  + 
  coord_fixed()

works

But this is not what I want. Rather, I want the y-value each pie chart is centered at to correspond to the 'yes_but_no' proportion for that specific pie chart

Solution

  • The problem arises because you recycled the same column for the y mapping & pie columns.

    The underlying code for geom_scatterpie does some data manipulations to pivot/gather (depending on the terminology of your tidyr package) the dataframe to long form, so the yes_but_no column no longer exists on its own by the time ggplot tries to plot the finalised data. The following would work:

    my_df$y <- my_df$yes_but_no # create a duplicate column

ggplot() + 
  geom_scatterpie(aes(x = day_in_july, y = y), 
                  data = my_df, 
                  cols = colnames(my_df)[2:3])

    For better appearance, you can apply equal coordinates while spacing out the y positions:

    ggplot() + 
  geom_scatterpie(aes(x = day_in_july, y = y*100), # space y positions out
                  data = my_df, 
                  cols = colnames(my_df)[2:3]) +
  scale_y_continuous(name = "yes but no",
                     labels = function(x) x/100) + # maintain original y labels
  coord_fixed()

    result