CODEWITHSUNDEEP
rggplot2
Alan
Alan

Reputation: 3

Ordering x-axis on correlation heat map using ggplot

I am trying to create a correlation heat map using ggplot, but I cannot seem to control the order of my variables on the x or y axis. Specifally, ggplot seems to try to order the variables sequentially, but only operating on the first digit. Here's a simple example of the problem

x1<-c(1,2,3,4,5,6,6,7,4,5,6,3,4,5,6,7,8,9)

x2<-c(1,2,3,3,5,4,6,7,4,4,6,3,4,5,6,10,8,9)

x3<-c(2,2,3,5,5,4,6,4,4,4,6,3,4,5,6,10,8,9)

x4<-c(1,2,3,5,5,4,6,4,4,4,4,4,4,5,6,10,8,9)

x5<-c(1,2,3,3,5,4,6,7,4,4,6,3,4,5,6,10,8,9)

x10<-c(1,1,1,1,1,1,6,1,1,1,1,1,4,5,1,1,1,1)

new=data.frame(x1)

new$x2=x2

new$x3=x3

new$x4=x4

new$x5=x5

new$x10=x10

keep=melt(cor(new))

ggplot(keep,aes(Var1,Var2),xlab=NULL,ylab=NULL) + geom_tile(aes(fill = value),colour = "white") + scale_fill_gradient(low = "white",high = "steelblue")+theme(axis.title.x=element_blank(),axis.title.y=element_blank())

If you run this code you'll see that the x-axis is ordered 

x1, x10, x2, x3, x4, x5

whereas I want it to read

x1, x2, x3, x4, x5, x10

Is there a way to specify this order?

Upvotes: 0

Views: 3187

Answers (4)

Isin Altinkaya
Isin Altinkaya

Reputation: 459

Based on @Troy's answer I wrote a small function:

reorder_version_sorting <- function(col,prefix){
  unique(col[order(as.numeric(gsub(prefix,"",col)))])
}

Usage:

ggplot +...+
      scale_x_discrete(limits=reorder_version_sorting(dd2$Rep,"rep"))

Before:

"rep0"  "rep10" "rep11" "rep12" "rep13" "rep14" "rep15" "rep16" "rep17" "rep18" "rep19" "rep1"  "rep2"  "rep3"  "rep4"  "rep5"  "rep6"  "rep7"  "rep8"  "rep9"

After:

"rep0"  "rep1"  "rep2"  "rep3"  "rep4"  "rep5"  "rep6"  "rep7"  "rep8"  "rep9"  "rep10" "rep11" "rep12" "rep13" "rep14" "rep15" "rep16" "rep17" "rep18" "rep19"

Upvotes: 0

Ben
Ben

Reputation: 42283

This works for me:

keep$Var1 <- factor(keep$Var1, levels = unique(keep$Var1), ordered = TRUE)
keep$Var2 <- factor(keep$Var2, levels = unique(keep$Var2), ordered = TRUE)

ggplot(keep,aes(Var1,Var2),xlab=NULL,ylab=NULL) + geom_tile(aes(fill = value),colour = "white") + scale_fill_gradient(low = "white",high = "steelblue")+theme(axis.title.x=element_blank(),axis.title.y=element_blank())

enter image description here

Upvotes: 1

Troy
Troy

Reputation: 8691

Try this if the padded zero solution doesn't fit for your purpose:

cust_breaks<-unique(keep$Var1[order(as.numeric(gsub("x","",keep$Var1)))])

ggplot(keep,aes(Var1,Var2),xlab=NULL,ylab=NULL) + 
  geom_tile(aes(fill = value),colour = "white") + 
  scale_fill_gradient(low = "white",high = "steelblue") + 
  theme(axis.title.x=element_blank(),axis.title.y=element_blank()) +
  scale_x_discrete(limits=cust_breaks) + 
  scale_y_discrete(limits=cust_breaks)

Upvotes: 0

John Paul
John Paul

Reputation: 12664

For this particular example, one option would be to name your variables x01, x02, etc and then they will order correctly.

Upvotes: 0

Related Questions