[daily-r]factor->characteristic->number conversion
I want to plot line graphs to show the y changes along with time series, just like in this post: http://stackoverflow.com/questions/12500218/ggplot2-line-plotting-with-time-series-and-multi-spline/12500368#comment16830065_12500368.
However, the x-axis treat the time points as five groups rather than numbers, as in the following figure.
This one is really what I am talking about:
==
Test raw data and R scripts to draw the first wrong figure are followed.
1 a 4.17125 41.33875 29.674375 8.551875 5.5 2 b 4.101875 29.49875 50.191875 13.780625 4.90375 3 c 3.1575 29.621875 78.411875 25.174375 7.8012
1 df <- read.delim("~/Desktop/df.b", header=F) 2 colnames(df)<-c("sample",0,15,30,60,120) 3 df2<-melt(df,id="sample") 4 ggplot(data = df2, aes(x=variable, y= value, group = sample, colour=sample)) + geom_line() + geom_point()
After the following conversion, the R script is right.
df2$variable <- as.numeric(as.character(df2$variable))
Factors look like character vectors but they are not! They are integer codes and a set of labels that go with the codes. Hence, as.numeric will simply return the integer codes. That's why I converted to character first. A more cryptic way to do it would be as.numeric(levels(df2$Time)[df2$Time])
