CODEWITHSUNDEEP
r
user3022875
user3022875

Reputation: 9018

R ggplot stacking not showing lines no chart

I want to plot the 4 columns of a matrix and have the x-axis say (1,2,3,4). I have provided a simple example you can run.

require("reshape")
require("ggplot2")

a<-rbind(.493,.537,.50,.462)
b<-rbind(.846,-.117,-.349,-.385)
c<-rbind(-.181,.0657,.135,-.719)
d<-rbind(-.09,.51,-.77,.34)

   me<- cbind(a,b,c,d)
data<-c(me[,1], me[,2], me[,3], me[,4])
description<-rep(    c("PC1","PC2","PC3","PC4"), each=NROW(me)  )
d<- cbind(data, description)
ax <- rep(    cbind(seq(from= 1, to = 4, by =1)), NROW(me)  )
 stacked <- data.frame(d , xaxis= ax )  
stacked

ggplot(data=stacked, aes( x = xaxis,  y=data, colour=description))   +   geom_line()

The problem is that there should be 4 lines showing on the chart but it is blank.

Any ideas?

Thank you.

Upvotes: 0

Views: 124

Answers (2)

Brian Diggs
Brian Diggs

Reputation: 58845

As I mentioned in my comment, you are converting your numeric values to character strings when you use cbind to create the matrix you call d. You are making it much harder than it is. Going back to your starting matrix me (simpler way specify the starting matrix rather than creating the steps you had; got it using dump("me", ""))

me <-
structure(c(0.493, 0.537, 0.5, 0.462, 0.846, -0.117, -0.349, 
-0.385, -0.181, 0.0657, 0.135, -0.719, -0.09, 0.51, -0.77, 0.34
), .Dim = c(4L, 4L))

You are using "PC1", "PC2", to identify the columns, so include that in the matrix directly

colnames(me) <- c("PC1", "PC2", "PC3", "PC4")

Now me looks like

> me
       PC1    PC2     PC3   PC4
[1,] 0.493  0.846 -0.1810 -0.09
[2,] 0.537 -0.117  0.0657  0.51
[3,] 0.500 -0.349  0.1350 -0.77
[4,] 0.462 -0.385 -0.7190  0.34

The rows don't have names, but are just indexed counts which match exactly what you want for xaxis later. Now just melt this:

library("reshape2")
stacked <- melt(me)

which looks like

> stacked
   Var1 Var2   value
1     1  PC1  0.4930
2     2  PC1  0.5370
3     3  PC1  0.5000
4     4  PC1  0.4620
5     1  PC2  0.8460
6     2  PC2 -0.1170
7     3  PC2 -0.3490
8     4  PC2 -0.3850
9     1  PC3 -0.1810
10    2  PC3  0.0657
11    3  PC3  0.1350
12    4  PC3 -0.7190
13    1  PC4 -0.0900
14    2  PC4  0.5100
15    3  PC4 -0.7700
16    4  PC4  0.3400

and the columns are of the appropriate type

> str(stacked)
'data.frame':   16 obs. of  3 variables:
 $ Var1 : int  1 2 3 4 1 2 3 4 1 2 ...
 $ Var2 : Factor w/ 4 levels "PC1","PC2","PC3",..: 1 1 1 1 2 2 2 2 3 3 ...
 $ value: num  0.493 0.537 0.5 0.462 0.846 -0.117 -0.349 -0.385 -0.181 0.0657 ...

The column names are different than what you had in your stacked, but they are all there. The graphing call is then just

ggplot(stacked, aes(x=Var1, y=value, colour=Var2)) + 
  geom_line()

enter image description here

Upvotes: 0

Ananta
Ananta

Reputation: 3711

your data variable is factor not numeric. 

ggplot(data=stacked, aes( x = xaxis,  y=as.numeric(as.character(data)), colour=description))   +   geom_line()

You are making dataframe hardway. I cleaned a little bit.

a<-c(.493,.537,.50,.462)
b<-c(.846,-.117,-.349,-.385)
c<-c(-.181,.0657,.135,-.719)
d<-c(-.09,.51,-.77,.34)

data<-c(a,b,c,d)
description<-rep(c("PC1","PC2","PC3","PC4"), each=4)
ax <- rep(cbind(seq(from= 1, to = 4, by =1)), 4  )
stacked <- data.frame(data, description, xaxis= ax)  
ggplot(data=stacked, aes( x = xaxis,  y=data, colour=description))+geom_line()

Upvotes: 1

Related Questions