1

With sample and code below, I'm trying to reproduce the effect of an image by adding geom_point for only type 3 data as shown in the end of this question:

library(data.table)
library(ggplot2)
df <- structure(list(date = c("2021-07-31", "2021-08-31", "2021-09-07", 
"2021-09-14", "2021-09-21", "2021-09-30", "2021-10-7", "2021-10-14", 
"2021-10-21", "2021-10-31", "2021-11-30", "2021-12-31", "2022-1-31", 
"2022-2-28"), value = c(190.3, 174.9, 163.2, 168.4, 168.6, 168.2, 
163.5, 161.6, 172.9, 166.5, 175.2, 197.7, 212.1, 177.9), type = c(1L, 
1L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 2L, 3L, 3L, 3L, 3L)), class = "data.frame", row.names = c(NA, 
-14L))
df$type <- as.factor(df$type)
df$date <- as.Date(df$date)
setDT(df, key = "date")
types.extend <- c(1)
new.values <- df[, .SD[1], by = .(type)][, type := shift(type, type = "lag")][type %in% types.extend,]
# bind the new value to your original data
df <- rbind(df, new.values)
ggplot(data = df, 
       aes(x=date, y=value, group=type, color = type, fill = type)) + 
  geom_area(alpha=0.4, position = "identity") +
  geom_point(size=2) +
  scale_color_manual(values=c("1" = "gray", "2" = "red", "3" = "blue"), 
                     aesthetics = c("color", "fill"), name = "type") +
  theme_bw()

Out:

enter image description here

As I add geom_point(size=2) to the code, it add points for all data points, how could I add geom_point only for data's type is 3? Sincere thanks.

enter image description here

ah bon
  • 9,293
  • 12
  • 65
  • 148

2 Answers2

3

You can use the data argument of geom_point to restrict the points being plotted. So replace your geom_point(size=2) call with geom_point(data=df[type==3], size=2)

(note this syntax only works because per the question df is a data.table, use df[df$type==3,] otherwise)

enter image description here

An alternative, if you didn't want the legend to include the points for the other two types would be to set the shape to NA for the types you didn't want to include:

 geom_point(size=2, aes(shape=type)) +
  scale_shape_manual(values=c("1"=NA, "2"=NA, "3"=19))+

enter image description here

ah bon
  • 9,293
  • 12
  • 65
  • 148
George Savva
  • 4,152
  • 1
  • 7
  • 21
  • Thanks, to go further, if I'd like to plot the last datapoint of type 2 + datapoints of type 3 as shown in the example figure? – ah bon Nov 16 '21 at 11:12
  • 1
    geom_point(data=df[type==3 | date==max(df[type==2]$date)],size=2, aes(shape=type)) + scale_shape_manual(values=c("1"=NA, "2"=19, "3"=19))+... – George Savva Nov 16 '21 at 12:35
1

Not sure if I understood your question but try :

ggplot(data = df, 
       aes(x=date, y=value, group=type, color = type, fill = type)) + 
  geom_area(alpha=0.4, position = "identity") +
  geom_point(data = filter(df, type == 3), aes(x=date, y=value, group=type), size = 2) +
  scale_color_manual(values=c("1" = "gray", "2" = "red", "3" = "blue"), 
                     aesthetics = c("color", "fill"), name = "type") +
  theme_bw()

Out:

enter image description here

ah bon
  • 9,293
  • 12
  • 65
  • 148
Flavien PETIT
  • 106
  • 7
  • Thanks a lot. To go further, if I'd like to plot the last datapoint of type 2 + datapoints of type 3 as shown in the example figure? – ah bon Nov 16 '21 at 11:12