0

I am trying to sort the graph by count, but I am not clearly getting how I can. Important thing is that 'topic' is a discrete variable and I'm just showing the count of each topic through the whole data set. This is the code I am using:

ggplot(data = dtd, aes(x = topic)) + geom_bar(fill = "lightblue", colour = "black") + coord_polar()

And this is the error when I try to reorder:

 ggplot(data = dtd, aes(x = reorder(topic))) + geom_bar(fill = "lightblue", colour = "black") + coord_polar()
Error in tapply(X = X, INDEX = x, FUN = FUN, ...) : 
  argument "X" is missing, with no default

It'd be a plus if someone might say how I could fill those sections with a gradient blue shade or something pretty.

MelaniaCB
  • 427
  • 5
  • 16

1 Answers1

1

The error is coming from reorder not having a second vector of values to reorder by. You could supply one by counting up your data ahead of passing to ggplot, then reordering it by the new variable:

library(dplyr)
library(ggplot2)

df <- tibble(Topic = c(rep("A", 55), rep("B", 22), rep("C", 40), rep("D", 100)))

df %>%
  group_by(Topic) %>% 
  summarise(freq = n()) %>% 
  ggplot(aes(x = reorder(Topic, freq), y = freq)) + 
  geom_bar(fill = "lightblue", colour = "black", stat = "identity") +
  coord_polar()

There may be smoother ways of doing it, but this keeps the reorder function you were using before.

Andy Baxter
  • 5,833
  • 1
  • 8
  • 22
  • Also, sorry not to have addressed the gradient fill question. Turns out, it's complicated https://stackoverflow.com/questions/53397131/gradient-fill-in-ggplot2 - you're probably best just saving it as a vector file and manually doing the gradient! – Andy Baxter Dec 03 '19 at 23:10