R help - Dec 2004 - Adding a polygon to a time series plot

Preamble: Sorry for being dense.

Now that that's done, here are my questions.

I want to put a polygon on a plot of a time series. I'm going to add lines
from a smooth.spline interpolation and other annotation to it. But here's
the general idea:

# Start code
     n <- 121
     dat <- rnorm(n)
     plot(dat, type="n")
     polygon(c(1:n,n:1), c(dat,c(rep(0,n))), col = "gray80", border =
NA)
# End code

What if dat is a time series?

# Start code that won't parse.
     dat.ts <- ts(rnorm(n), start = 1980, frequency = 12)
     plot(dat.ts, type="n")
     polygon(...) # How do I specify x and y for the time series?
                  # Is it easier to skip the time series and add time as an
axis
# End code that won't parse

Here's a work around that is cumbersome.

# Start code
     dat.ts <- ts(rnorm(n), start = 1980, frequency = 12)
     dat <- as.vector(dat.ts)
     plot(1:n, dat, type="n", xaxt = "n", axes = FALSE, xlab
= "", ylab "")
     polygon(c(1:n,n:1), c(dat,c(rep(0,n))), col = "gray80", border =
NA)
     axis(1, at = seq(1, n, by = 12), labels = c(1980:1990))
# End code

Can I use polygon with a time series object? Or is it too awkward?

Thanks, Andy

> From: Andy Bunn
> 
> Preamble: Sorry for being dense.
> 
> Now that that's done, here are my questions.
> 
> I want to put a polygon on a plot of a time series. I'm going 
> to add lines
> from a smooth.spline interpolation and other annotation to 
Please note: smooth.spline() is for _smoothing_.  If you really want
interpolation, you should use spline() or splinefun().
> it. But here's
> the general idea:
> 
> # Start code
>      n <- 121
>      dat <- rnorm(n)
>      plot(dat, type="n")
>      polygon(c(1:n,n:1), c(dat,c(rep(0,n))), col = "gray80", 
> border = NA)
> # End code
> 
> What if dat is a time series?
> 
> # Start code that won't parse.
>      dat.ts <- ts(rnorm(n), start = 1980, frequency = 12)
>      plot(dat.ts, type="n")
>      polygon(...) # How do I specify x and y for the time series?
>                   # Is it easier to skip the time series and 
> add time as an
> axis
> # End code that won't parse
AFAICS, the plot will have the x-axis range equal to the time you specified,
so replace 1 with start(dat.ts) and n with end(dat.ts) ought to work, no?

HTH,
Andy

 
> Here's a work around that is cumbersome.
> 
> # Start code
>      dat.ts <- ts(rnorm(n), start = 1980, frequency = 12)
>      dat <- as.vector(dat.ts)
>      plot(1:n, dat, type="n", xaxt = "n", axes = FALSE,
xlab
> = "", ylab > "")
>      polygon(c(1:n,n:1), c(dat,c(rep(0,n))), col = "gray80", 
> border = NA)
>      axis(1, at = seq(1, n, by = 12), labels = c(1980:1990))
> # End code
> 
> Can I use polygon with a time series object? Or is it too awkward?
> 
> Thanks, Andy
> 
> ______________________________________________
> R-help at stat.math.ethz.ch mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide! 
> http://www.R-project.org/posting-guide.html
> 
>