I'm trying to plot a data set containing over 2000 samples in a stacked bar chart format, with each sample (represented by "SampleID") on the x-axis and 6 measurement values on the y-axis (Measurement1-6
). I want the samples to be displayed/ordered by a the following order of measurement variables: Measurement4
, 1, 5, 2, 3, and 6, and from highest to lowest measurement value. Below is a subset of 15 samples as an example of what I'm working with, which I'll refer to as the "dummy_set" data frame:
SampleID Measurement1 Measurement2 Measurement3 Measurement4 Measurement5 Measurement6
1 A 0.05 0.00 0.95 0.00 0.0 0.00
2 B 0.00 0.00 0.43 0.56 0.0 0.01
3 C 0.64 0.36 0.00 0.00 0.0 0.00
4 D 0.00 0.82 0.18 0.00 0.0 0.00
5 E 0.00 0.60 0.00 0.40 0.0 0.00
6 F 0.80 0.00 0.00 0.20 0.0 0.00
7 G 0.00 0.00 0.00 1.00 0.0 0.00
8 H 0.00 0.00 0.00 1.00 0.0 0.00
9 I 0.00 0.00 1.00 0.00 0.0 0.00
10 J 0.00 0.00 1.00 0.00 0.0 0.00
11 K 0.25 0.00 0.00 0.45 0.3 0.00
12 L 0.10 0.00 0.00 0.10 0.8 0.00
13 M 0.19 0.10 0.00 0.70 0.0 0.01
14 N 0.90 0.00 0.00 0.10 0.0 0.00
15 O 0.00 0.10 0.40 0.00 0.5 0.00
Here's the basics of what I've done:
Melt the data set:
melt_dummy_set <- melt(dummy_set, id.var = "SampleID")
Where the melted data set looks like this:
head(melt_dummy_set)
SampleID variable value
1 A Measurement1 0.05
2 B Measurement1 0.00
3 C Measurement1 0.64
4 D Measurement1 0.00
5 E Measurement1 0.00
6 F Measurement1 0.80
Plot the melted data set using ggplot() and geom_bar():
ggplot(melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
As you can see, the samples are plotted in the original order that they were listed (A-O). However, I want them to be plotted in the following order: G, H, M, B, K, N, F, C, L, O, D, E, I, J, and A.
Based on other similar questions, I've gathered that I need to relevel/re-establish the factors in the order I want. Here is what I've tried so far:
#Attempt 1
reordered_melt_dummy_set <- transform(melt_dummy_set, variable = reorder(variable, -value))
ggplot(reordered_melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
Attempt 1 results
#Attempt 2
copy_melt_dummy_set <- melt_dummy_set
copy_melt_dummy_set$variable <- factor(copy_melt_dummy_set$variable, levels = c("Measurement4", "Measurement5", "Measurement1", "Measurement2", "Measurement3", "Measurement6"))
ggplot(copy_melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
Attempt 2 results
My 3rd attempt resulted in multiple errors (denoted in "##" immediately after the line of code)
#Attempt 3
copy2_melt_dummy_set <- melt_dummy_set
copy2_melt_dummy_set$SampleID <- factor(copy2_melt_dummy_set$SampleID, levels = copy2_melt_dummy_set[order(-copy2_melt_dummy_set$value), "variable"])
##Error in `levels<-`(`*tmp*`, value = as.character(levels)) : factor level [2] is duplicated
copy2_melt_dummy_set$variable <- factor(copy2_melt_dummy_set$variable, levels = copy2_melt_dummy_set[order(copy2_melt_dummy_set$value), "variable"])
## Error in `levels<-`(`*tmp*`, value = as.character(levels)) : factor level [2] is duplicated
copy2_melt_dummy_set$SampleID <- factor(copy2_melt_dummy_set$SampleID, levels = copy2_melt_dummy_set[order(-copy2_melt_dummy_set$variable), "SampleID"])
## Error in `levels<-`(`*tmp*`, value = as.character(levels)) : factor level [16] is duplicated
## In addition: Warning message: In Ops.factor(copy2_melt_dummy_set$variable) : ‘-’ not meaningful for factors
copy2_melt_dummy_set$SampleID <- factor(copy2_melt_dummy_set$SampleID, levels = copy2_melt_dummy_set[order(-copy2_melt_dummy_set$value), "SampleID"])
## Error in `levels<-`(`*tmp*`, value = as.character(levels)) : factor level [16] is duplicated
copy2_melt_dummy_set$SampleID <- factor(copy2_melt_dummy_set$SampleID, levels = copy2_melt_dummy_set[order(-copy2_melt_dummy_set$value), "value"])
## Error in `levels<-`(`*tmp*`, value = as.character(levels)) : factor level [2] is duplicated
#Attempt 4
copy3_melt_dummy_set <- melt_dummy_set[order(melt_dummy_set$variable, -melt_dummy_set$value), ]
head(copy3_melt_dummy_set)
ggplot(copy3_melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
Attempt 4 results
#Attempt 5
ggplot(melt_dummy_set[order(melt_dummy_set$variable, -melt_dummy_set$value), ], aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
Attempt 5 results
#Attempt 6
new_melt_dummy_set <- within(melt_dummy_set,
variable <- factor(variable, levels = names(sort(table(variable), decreasing = TRUE))))
ggplot(new_melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
Attempt 6 results
#Attempt 7
copy4_melt_dummy_set <- melt_dummy_set
custom_leveling <- unique(copy4_melt_dummy_set$variable)
copy4_melt_dummy_set$variable <- factor(copy4_melt_dummy_set$variable, level = custom_leveling)
ggplot(copy4_melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
Attempt 7 results
In all cases I can't get the actual samples on the x-axis to be reorganized. I feel there's probably a simple fix for this, but I can't figure out what I'm doing wrong. Any suggestions?
Edited
In response to the possible duplicate comment, I tried applying the codes/solutions from Order Bars in ggplot2 bar graph and they did not produce the plot in the desired order that I wanted. See below for the codes I tried:
#First solution
new_melt_dummy_set <- within(melt_dummy_set,
variable <- factor(variable, levels = names(sort(table(variable), decreasing = TRUE))))
ggplot(new_melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity")
#Second solution
ggplot(melt_dummy_set, aes(x = reorder(SampleID, variable, function(x)-length(x)), y = value, fill = variable)) + geom_bar(stat = "identity")
ggplot(melt_dummy_set, aes(x = reorder(variable, SampleID, function(x)-length(x)), y = value, fill = variable)) + geom_bar(stat = "identity")
#Third solution
ordered_measurements <- c("Measurement4", "Measurement1", "Measurement5", "Measurement2", "Measurement3", "Measurement6")
ggplot(melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
scale_x_discrete(limits = ordered_measurements)
#Fourth solution
ggplot(melt_dummy_set, aes(x = reorder(SampleID, -table(variable)[variable]), y = value, fill = variable)) + geom_bar(stat = "identity")
require(forcats)
ggplot(melt_dummy_set, aes(x = SampleID, fill = fct_infreq(variable), y = value)) + geom_bar(stat = "identity")
ggplot(melt_dummy_set, aes(x = fct_infreq(variable))) + geom_bar(stat = "identity")
#Fifth solution
library(tidyverse)
library(forcats)
melt_dummy_set %>%
mutate(variable = fct_reorder(variable, value, .desc = TRUE)) %>%
ggplot(aes(x = SampleID, y = value, fill = variable)) + geom_bar(stat = 'identity')
#Sixth solution
library(dplyr)
melt_dummy_set %>%
group_by(variable) %>%
summarize(counts = n()) %>%
arrange(-counts) %>%
mutate(SampleID = factor(SampleID, variable)) %>%
ggplot(aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity")
melt_dummy_set %>%
group_by(SampleID) %>%
summarize(counts = n()) %>%
arrange(-counts) %>%
mutate(SampleID = factor(SampleID, value)) %>%
ggplot(aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity")
#Seventh solution
new_meltedDummy_set <- transform(melt_dummy_set,
variable = ordered(variable, levels = names(sort(-table(variable)))))
ggplot(new_meltedDummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity")
Is this what you where going for? I think you were close. Instead of turning the Measurement
variable column into a factor you need to order the SampleID
column based on the order of Measurement
values. This is what happens in the line where sample_order
is calculated:
library(tidyverse)
dummy_set <- tibble(
SampleID = c("A", "B", "C", "D", "E", "F", "G", "H", "I", "J", "K", "L", "M", "N", "O"),
Measurement1 = c(0.05, 0, 0.64, 0, 0, 0.8, 0, 0, 0, 0, 0.25, 0.1, 0.19, 0.9, 0),
Measurement2 = c(0, 0, 0.36, 0.82, 0.6, 0, 0, 0, 0, 0, 0, 0, 0.1, 0, 0.1),
Measurement3 = c(0.95, 0.43, 0, 0.18, 0, 0, 0, 0, 1, 1, 0, 0, 0, 0, 0.4),
Measurement4 = c(0, 0.56, 0, 0, 0.4, 0.2, 1, 1, 0, 0, 0.45, 0.1, 0.7, 0.1, 0),
Measurement5 = c(0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.3, 0.8, 0, 0, 0.5),
Measurement6 = c(0, 0.01, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0.01, 0, 0)
)
sample_order <- dummy_set %>%
arrange(desc(Measurement4), desc(Measurement1), desc(Measurement5), desc(Measurement2), desc(Measurement3), desc(Measurement6)) %>%
pull(SampleID)
melt_dummy_set <- dummy_set %>%
gather(variable, value, -SampleID)
reordered_melt_dummy_set <- melt_dummy_set %>%
mutate(SampleID = factor(SampleID, levels = sample_order))
plot_ordered <- ggplot(reordered_melt_dummy_set, aes(x = SampleID, y = value, fill = variable)) +
geom_bar(stat = "identity") +
scale_y_continuous(expand = c(0,0)) +
theme(axis.ticks.x = element_blank(), panel.grid = element_blank(), axis.line = element_line(color = "black"), panel.border = element_blank(), panel.background = element_blank())
plot_ordered
Created on 2019-07-26 by the reprex package (v0.3.0)