CODEWITHSUNDEEP
rggplot2
gantonioid
gantonioid

Reputation: 457

Assign point color depending on data.frame column value R

this is my first question on SO, I hope someone can help me answer it.

I'm reading data from a csv with R with data<-read.csv("/data.csv") and get something like:

Group    x   y  size    Color
Medium   1   2  2000    yellow
Small   -1   2  1000    red
Large    2  -1  4000    green
Other   -1  -1  2500    blue

Each group color may vary, they are assigned by a formula when the csv file is generated, but those are all the possible colors (the number of groups may also vary).

I've been trying to use ggplot() like so:

data<-read.csv("data.csv")
xlim<-max(c(abs(min(data$x)),abs(max(data$x))))
ylim<-max(c(abs(min(data$y)),abs(max(data$y))))
data$Color<-as.character(data$Color)
print(data)
ggplot(data, aes(x = x, y = y, label = Group)) +
geom_point(aes(size = size, colour = Group), show.legend = TRUE) +
scale_color_manual(values=c(data$Color)) +
geom_text(size = 4) +
scale_size(range = c(5,15)) +
scale_x_continuous(name="x", limits=c(xlim*-1-1,xlim+1))+
scale_y_continuous(name="y", limits=c(ylim*-1-1,ylim+1))+
theme_bw()

Everything is correct except for the colors

I noticed the legend at the right orders the Groups alphabetically (Large, Medium, Other, Small), but the colors stay in the csv file order.

Here is a screenshot of the plot.

enter image description here

Can anyone tell me what's missing in my code to fix this? other approaches to achieve the same result are welcome.

Upvotes: 24

Views: 36278

Answers (2)

ScottyJ
ScottyJ

Reputation: 1057

A Slightly Better Solution...

I had never heard of R back when this question was answered by @scoa, and I don't know if my solution was available, but you can do what the OP asks with slightly less work using scale_color_identity().

library(tidyverse)

data <- tribble(
  ~Group,~x,~y,~size,~Color,
  "Medium",1,2,2000,"yellow",
  "Small",-1, 2,1000,"red",
  "Large",2,-1,4000,"green",
  "Other",-1,-1,2500,"blue")

xlim<-max(c(abs(min(data$x)),abs(max(data$x))))
ylim<-max(c(abs(min(data$y)),abs(max(data$y))))

ggplot(data, aes(x = x, y = y, label = Group)) +
  geom_point(aes(size = size, colour = Color), show.legend = TRUE) +   # Set aes(colour = Color) (the column in the dataframe)
  scale_color_identity() +  # This tells ggplot to use the values explicit in the 'Color' column
  geom_text(size = 4) +
  scale_size(range = c(5,15)) +
  scale_x_continuous(name="x", limits=c(xlim*-1-1,xlim+1))+
  scale_y_continuous(name="y", limits=c(ylim*-1-1,ylim+1))+
  theme_bw()

enter image description here

scale_color_identity()

By using this, you don't need to create the separate named vector that you do with scale_color_manual() and you can directly use the 'Color' column (note the change in geom_point(aes(colour = Group,... to geom_point(aes(colour = Color,...!!!).

Upvotes: 13

scoa
scoa

Reputation: 19867

One way to do this, as suggested by help("scale_colour_manual") is to use a named character vector:

col <- as.character(data$Color)
names(col) <- as.character(data$Group)

And then map the values argument of the scale to this vector

# just showing the relevant line
scale_color_manual(values=col) +

full code

xlim<-max(c(abs(min(data$x)),abs(max(data$x))))
ylim<-max(c(abs(min(data$y)),abs(max(data$y))))

col <- as.character(data$Color)
names(col) <- as.character(data$Group)

ggplot(data, aes(x = x, y = y, label = Group)) +
  geom_point(aes(size = size, colour = Group), show.legend = TRUE) +
  scale_color_manual(values=col) +
  geom_text(size = 4) +
  scale_size(range = c(5,15)) +
  scale_x_continuous(name="x", limits=c(xlim*-1-1,xlim+1))+
  scale_y_continuous(name="y", limits=c(ylim*-1-1,ylim+1))+
  theme_bw()

Ouput:

enter image description here

Data

data <- read.table("Group    x   y  size    Color
Medium   1   2  2000    yellow
Small   -1   2  1000    red
Large    2  -1  4000    green
Other   -1  -1  2500    blue",head=TRUE)

Upvotes: 27

Related Questions