Question

This question’s theme is simple but drives me crazy:
1. how to use melt()
2. how to deal with multi-lines in single one image?

Here is my raw data:

a   4.17125 41.33875    29.674375   8.551875    5.5
b   4.101875    29.49875    50.191875   13.780625   4.90375
c   3.1575  29.621875   78.411875   25.174375   7.8012

Q1:
I’ve learn from this post Plotting two variables as lines using ggplot2 on the same graph to know how to draw the multi-lines for multi-variables, just like this:

The following codes can get the above plot. However, the x-axis is indeed time-series.

df <- read.delim("~/Desktop/df.b", header=F)
colnames(df)<-c("sample",0,15,30,60,120)
df2<-melt(df,id="sample")
ggplot(data = df2, aes(x=variable, y= value, group = sample, colour=sample)) + geom_line() + geom_point()

I wish it could treat 0 15 30 60 120 as real number to show the time series, rather than name_characteristics. Even having tried this, I failed.

row.names(df)<-df$sample
df<-df[,-1]
df<-as.matrix(df)
df2 <- data.frame(sample = factor(rep(row.names(df),each=5)), Time = factor(rep(c(0,15,30,60,120),3)),Values = c(df[1,],df[2,],df[3,]))
ggplot(data = df2, aes(x=Time, y= Values, group = sample, colour=sample)) 
        + geom_line() 
        + geom_point()

Loooooooooking forward to your help.

Q2:
I’ve learnt that the following script can add the spline() function for single one line, what about I wish to apply spline() for all the three lines in single one image?

n <-10
d <- data.frame(x =1:n, y = rnorm(n))
ggplot(d,aes(x,y))+ geom_point()+geom_line(data=data.frame(spline(d, n=n*10)))

Answer 1

Your variable column is a factor (you can verify by calling str(df2)). Just convert it back to numeric:

df2$variable <- as.numeric(as.character(df2$variable))

For your other question, you might want to stick with using geom_smooth or stat_smooth, something like this:

p <- ggplot(data = df2, aes(x=variable, y= value, group = sample, colour=sample)) + 
      geom_line() + 
      geom_point()

library(splines)
p + geom_smooth(aes(group = sample),method = "lm",formula = y~bs(x),se = FALSE)

which gives me something like this: