I'm new to R.
What I'm trying to do is to decompose a time series, which is given in hourly resolution.
My data looks like this:
Time traffic
6/7/2005 7:00 56718587433
6/7/2005 8:00 76456162968
6/7/2005 9:00 82534038485
6/7/2005 10:00 88796995092
...
7/28/2005 10:00 51528036132
7/28/2005 11:00 69610584123
7/28/2005 12:00 76364975533
7/28/2005 13:00 81281257078
In order to decompose my data to trend, seasonal and stationary components, this is what I tried:
library(xts)
library(forecast)
data<-read.csv("my_file.csv")
data<-ts(data[,2],frequency = 24*365, start=c(2005,6,7,7))
decompose(data)
but this is the error I'm getting:
Error in decompose(data): time series has no or less than 2 period
What am I doing wrong?
When you use frequency = 24*365 you are signaling that you want the unit of time to be years. Since you have less than two months of data, you do not have 2 periods (years). Try using days as your basic unit instead - i.e. try frequency = 24