Sorting ggplot2 geom_posterior plot

Question

I have a dataset with 3 sets distribution each with 100 observation, I would like to plot these using ggplot posterior plot.

Data:

Name Value
A1    1
A1    1.2
A1    0.5
A2    5
A2    3
A2    -1
A3    5
A3    2
A3    1

Problem:

I am unable to sort the value by the mean value, the expected result should be "A2, A3, A1" in the plot.

Attempted solution:

ggplot(value1, aes(reorder(mean(value)), x = value, y = name))+
  geom_posterior() +
  xlab("value")  + theme_bw()

Full Code with Data:

library("tidyverse")
library("ggdistribute") #plot geom_posterior

value1 <- rnorm(n=100, mean=1, sd=1) %>% as.data.frame()
value2 <- rnorm(n=100, mean=3, sd=1) %>% as.data.frame()
value3 <- rnorm(n=100, mean=1, sd=2) %>% as.data.frame()

value1$name <- "A1"
value2$name <- "A2"
value3$name <- "A3"

value1 <- rbind(value1,value2,value3) 
colnames(value1) <- c("value","name")

ggplot(value1, aes(reorder(mean(value)), x = value, y = name))+
  geom_posterior() +
  xlab("value")  + theme_bw()

Result so far:

Answer 1

This may be a bit of a workaround - if you change name to an ordered factor and provide the levels in the order you'd like the graphs to be shown, then ggplot uses the factor levels for ordering:

my_levels <- value1 %>%
    group_by(name) %>%
    summarize(mean = mean(value)) %>% 
    arrange(mean) %>%
    select(name)

value1$name <-factor(value1$name, levels = my_levels$name, ordered = TRUE)

ggplot(value1, aes(reorder(mean(value)), x = value, y = name))+
  geom_posterior() +
  xlab("value")  + theme_bw()

Answer 2

You can reorder by mean:

reorder(name, value, mean)

specifying the vector whose levels will be reordered, a vector whose subset of values determines the order, and a function to apply to subsets (in this case, mean).

library(tidyverse)
library(ggdistribute)

set.seed(123)

value1 <- rnorm(n=100, mean=1, sd=1) %>% as.data.frame()
value2 <- rnorm(n=100, mean=3, sd=1) %>% as.data.frame()
value3 <- rnorm(n=100, mean=1, sd=2) %>% as.data.frame()

value1$name <- "A1"
value2$name <- "A2"
value3$name <- "A3"

value1 <- rbind(value1,value2,value3) 
colnames(value1) <- c("value","name")

ggplot(value1, aes(x = value, y = reorder(name, value, mean)))+
  geom_posterior() +
  xlab("value")  + theme_bw()

Plot