r pie chart labels overlap ggplot2

Question

I'm trying to make a pie chart with several slices, and many of them have low values. The problem is that when I make the chart most of the labels overlap each other.

The graphic is this:

The data:

           Descripcion  Freq
               Sumarios   17
    Previsiones Legales   34
          Multas SICORE   19
           Multas ANSeS    7
            Multas AFIP    5
  Gastos Corresponsalía   22
      Faltantes de Caja  470
    Cargos Jubilaciones 2185
            ATM Fraudes   10
        ATM Diferencias  201

And the code:

#armo el grafico
pmas <- ggplot(cant_masivos_trim, aes(x=1, y=Freq, fill=Descripcion)) +
        geom_bar(stat="identity") +
        ggtitle(paste("Cantidad de Reportes - Carga Masiva"))
pmas <- pmas + coord_polar(theta='y')
pmas <- ggplot(cant_masivos_trim, aes(x=1, Freq, fill=Descripcion)) +
        ggtitle(paste("Cantidad de Reportes - Carga Masiva")) +
        coord_polar(theta='y')
pmas <- pmas + geom_bar(stat="identity", color='black') + guides(fill=guide_legend

(override.aes=list(colour=NA)))
pmas <- pmas + theme(axis.ticks=element_blank(),  # the axis ticks
          axis.title=element_blank(),  # the axis labels
          axis.text.y=element_blank()) # the 0.75, 1.00, 1.25 labels.
y.breaks <- cumsum(cant_masivos_trim$Freq) - cant_masivos_trim$Freq/2
pmas <- pmas +
    # prettiness: make the labels black
    theme(axis.text.x=element_text(color='black')) +
    scale_y_continuous(
        breaks=y.breaks,   # where to place the labels
        labels= (paste(cant_masivos_trim$Freq, percent(cant_masivos_trim$Freq/sum (cant_masivos_trim$Freq)), sep='\n'))) # the labels

I try to find a solution here, but have no luck. Does anybody have an idea?

Answer 1

Here is an attempt using ggrepel. The result for the pie chart is not really pretty, but I can't improve it. And afterwards, I provide another solution without pie charts at all.

library(ggplot2)
library(tibble)
library(scales)
library(ggrepel)
library(forcats)

df <- tribble(
  ~Descripcion,  ~Freq,
   "Sumarios",   17,
   "Previsiones Legales",   34,
   "Multas SICORE",   19,
   "Multas ANSeS",    7,
   "Multas AFIP",    5,
   "Gastos Corresponsalía",   22,
   "Faltantes de Caja",  470,
   "Cargos Jubilaciones", 2185,
   "ATM Fraudes",   10,
   "ATM Diferencias",  201)

I change df$Descripcionto a factor, and ordered by df$Freq, using forcats::fct_reorder. And then I change the order in the data frame, so the function to position the labels works correctly.

df$Descripcion <- fct_reorder(df$Descripcion, df$Freq)

df <- df[order(df$Freq, decreasing = TRUE), ]
df
# A tibble: 10 × 2
#               Descripcion  Freq
#                   <fctr> <dbl>
#  1               Sumarios    17
#  2    Previsiones Legales    34
#  3          Multas SICORE    19
#  4           Multas ANSeS     7
#  5            Multas AFIP     5
#  6  Gastos Corresponsalía    22
#  7      Faltantes de Caja   470
#  8    Cargos Jubilaciones  2185
#  9            ATM Fraudes    10
# 10        ATM Diferencias   201

I then define another data frame to place the labels. I chose the x.breaks through trial and error.

my_labels <- tibble(x.breaks = seq(1, 1.5, length.out = 10),
                    y.breaks = cumsum(df$Freq) - df$Freq/2,
                    labels = paste(df$Freq, percent(df$Freq/sum (df$Freq)), sep='\n'),
                    Descripcion = df$Descripcion)

And then the plot (note that I changed the theme(axis.x.text) to element_blank() as I add the labels through geom_label_repel() now)

pmas <- ggplot(df, aes(x = 1, y = Freq, fill = Descripcion)) +
  ggtitle(paste("Cantidad de Reportes - Carga Masiva")) +
  geom_bar(stat="identity", color='black') + 
  coord_polar(theta='y') + 
  guides(fill=guide_legend(override.aes=list(colour=NA)))+ 
  theme(axis.ticks=element_blank(),  # the axis ticks
        axis.title=element_blank(),  # the axis labels
        axis.text.y=element_blank(), # the 0.75, 1.00, 1.25 labels.
        axis.text.x = element_blank(), 
        panel.grid = element_blank()) +
  scale_fill_brewer(palette = "Set3", direction = -1)+
  geom_label_repel(data = my_labels, aes(x = x.breaks, y = y.breaks, 
                                        label = labels, fill = Descripcion),
                   label.padding = unit(0.1, "lines"),
                   size = 2,
                   show.legend = FALSE,
                   inherit.aes = FALSE)

pmas

Here is another version of the plot, where you do not need to provide another data frame for the labels. I chose to put the labels before the bars, but it is up to you. Note the expand_limits(y = -150) to ensure that the label is visible, and the coord_flip() so as the labels are more readable. I also use geom_col() in place of geom_bar(stat = "identity").

pmas2 <- ggplot(data = df, aes(x = Descripcion, y = Freq)) +
  geom_col(aes(fill = Descripcion) , show.legend = FALSE) +
  ggtitle(paste("Cantidad de Reportes - Carga Masiva")) +
  coord_flip() +
  geom_label(aes(label = paste(df$Freq, percent(df$Freq/sum(df$Freq)), sep = "\n"),
                y = -150, fill = Descripcion),
             show.legend = FALSE,
             size = 3, label.padding = unit(0.1, "lines")) +
  expand_limits(y = -150) +
  scale_fill_brewer(palette = "Set3", direction = -1) 

pmas2