[R-sig-finance] average values for time series

[Dirk Eddelbuettel]

On 27 April 2005 at 06:37, Jonathan Q. wrote:
| New to R, currently use FAME.  One of the things I like about FAME is
| that if I have a series, say the near month for a commodity, I can set
| the series property so when i view weekly data, the data is the
| average of the daily data.  alternatively, for other series like stock

There are probably several ways to do that. 

| prices I can have the series always display the last value (i.e.,
| monthly would take last value in month vs the average).  also able to

The its package does that pretty well:

> library(its)				## verbose output suppressed here
> IBM<-priceIts("IBM",quote="Close")	## fetches from Yahoo!
> head(IBM)
           IBM Close
1991-01-02    112.12
1991-01-03    112.50
1991-01-04    112.12
1991-01-07    110.25
1991-01-08    109.00
1991-01-09    106.87
> head(extractIts(IBM, find="last", period="week"))
           IBM Close
1991-01-04    112.12
1991-01-11    108.12
1991-01-18    117.62
1991-01-25    122.62
1991-02-01    126.87
1991-02-08    129.50
> head(extractIts(IBM, find="last", period="month"))
           IBM Close
1991-01-31    126.75
1991-02-28    128.75
1991-03-28    113.87
1991-04-30    103.00
1991-05-31    106.12
1991-06-28     97.12

I just taught myself how to the other part with zoo -- pretty slick:

> library(zoo)				## once again verbose output nixed
> IBMzoo <- zoo(IBM)			## turn its object into zoo object
> head(aggregate(IBMzoo, format(dates(IBM), "%Y-%m"), mean))

1991-01 115.4127
1991-02 132.2053
1991-03 123.3240
1991-04 109.6800
1991-05 104.2241
1991-06 100.8970
>

This runs aggregate() over the zoo() object where the on-the-fly factor is
computed from the underlying POSIXt date representation, and the aggregating
function, here mean, is then passed over each unique factor set.

Hope this helps,  Dirk

-- 
Better to have an approximate answer to the right question than a precise 
answer to the wrong question.  --  John Tukey as quoted by John Chambers