Some basic time series questions

classic Classic list List threaded Threaded
5 messages Options
Reply | Threaded
Open this post in threaded view
|

Some basic time series questions

Hall, Mark
Hi,

As I sit and learn how to work with time series, I've run into a problem
that is eluding a quick easy answer (and googling for answers seems to
really slow the process...)

Question #1--
In a simple example on R 3.3.1 (sorry my employer hasn't upgraded to 3.3.2
yet):

x=rnorm(26,0,1)
x.ts<-ts(x,start=c(2014,9),frequency=12)

inputting x.ts at the prompt gives me a table with the rooms denoted by
year and columns denoted by months and everything lines up wonderfully.

Now my problem comes when I type at the prompt

plot(x.ts)  or
plot(x.ts, xlab="") or
plot.ts(x.ts,xlab="")

I get a plot of the values, but my x-axis labels are 2015.0, 2015.5,
2016.0, and 2016.5 .  January 2015 is coming out as 2015.0...

Is there a way of getting a more intelligible x-axis labeling?  Even 2015.1
for Janaury, etc. would work, or even getting an index (either Septemebr
2014 representing 0 or 1 and it incrementally increasing each month).

Question #2--
If I have a time series of decadal events, how best should I set the
frequency.  It is historical data, in the form of say AD 610-619 5 events,
AD 620-629 7 events, etc.

Sorry for such a basic questions.  Any advice would be appreciated.

Thanks in advance, MEH



Mark E. Hall, PhD
Assistant Field Manager
Black Rock Field Office
Winnemucca District Office
775-623-1529.

        [[alternative HTML version deleted]]

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: Some basic time series questions

Paul Gilbert-2

On 11/17/2016 06:00 AM, [hidden email] wrote:

> Hi,
>
> As I sit and learn how to work with time series, I've run into a problem
> that is eluding a quick easy answer (and googling for answers seems to
> really slow the process...)
>
> Question #1--
> In a simple example on R 3.3.1 (sorry my employer hasn't upgraded to 3.3.2
> yet):
>
> x=rnorm(26,0,1)
> x.ts<-ts(x,start=c(2014,9),frequency=12)
>
> inputting x.ts at the prompt gives me a table with the rooms denoted by
> year and columns denoted by months and everything lines up wonderfully.
>
> Now my problem comes when I type at the prompt
>
> plot(x.ts)  or
> plot(x.ts, xlab="") or
> plot.ts(x.ts,xlab="")
>
> I get a plot of the values, but my x-axis labels are 2015.0, 2015.5,
> 2016.0, and 2016.5 .  January 2015 is coming out as 2015.0...
>
> Is there a way of getting a more intelligible x-axis labeling?  Even 2015.1
> for Janaury, etc. would work, or even getting an index (either Septemebr
> 2014 representing 0 or 1 and it incrementally increasing each month).

There are several ways to do this. Having some bias, I would do

   require(tfplot)
   tfplot(x.ts)

>
> Question #2--
> If I have a time series of decadal events, how best should I set the
> frequency.  It is historical data, in the form of say AD 610-619 5 events,
> AD 620-629 7 events, etc.

For anything other than annual, quarterly, and monthly data you probably
should consider zoo. (There are other options, but I think zoo is the
most widely used for some time now.) Guessing a bit how you think of
this data, I would say you want the date index to be the first year of
the decade. To illustrate, I can generate a hundred decades of random
data and index it thus:

require(zoo)
x <- zoo(round(10 * runif(100)) , order.by= 10 * 61:160)

Paul

>
> Sorry for such a basic questions.  Any advice would be appreciated.
>
> Thanks in advance, MEH
>
>
>
> Mark E. Hall, PhD
> Assistant Field Manager
> Black Rock Field Office
> Winnemucca District Office
> 775-623-1529.

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: require vs library ( Some basic time series questions)

Jeff Newmiller
> require(tfplot)
> tfplot(x.ts)

Would just like to point out that require() should not be treated as interchangeable with library(). The former returns a logical status indicating success or failure, while the latter throws an error if it falls.  You should reserve use of require() for cases when you are implementing an alternative path of execution for failure, and in nearly all usual cases use the library() function instead so hapless users of your script don't have to sift through all the subsequent errors to figure out the problem.
--
Sent from my phone. Please excuse my brevity.

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: require vs library ( Some basic time series questions)

Paul Gilbert-2


On 11/17/2016 04:49 PM, Jeff Newmiller wrote:

>> require(tfplot)
>> tfplot(x.ts)
>
> Would just like to point out that require() should not be treated as
> interchangeable with library(). The former returns a logical status
> indicating success or failure, while the latter throws an error if it
> falls.  You should reserve use of require() for cases when you are
> implementing an alternative path of execution for failure, and in
> nearly all usual cases use the library() function instead so hapless
> users of your script don't have to sift through all the subsequent
> errors to figure out the problem.
>
Mea culpa.  Force of habit from usually writing with an alternative path
of execution. In this example I should have used library(), especially
since 'tfplot' may not have been installed on the user's system and so
the library() error message would be more explicit than the warning from
require(). But "in nearly all usual cases" seems a bit strong. It
implies R users don't usually program much and thus do not implement an
alternative path of execution for failure. (Some of us consider that the
most usual case.)

Paul

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
Reply | Threaded
Open this post in threaded view
|

Re: require vs library ( Some basic time series questions)

Jeff Newmiller
In reply to this post by Jeff Newmiller
Perhaps more people write end-user-ready applications in R than I am aware of and my bias is too strong.  For working at the console I prefer not to have my scripts installing packages on their own (one possible alternative execution path), and it is too much trouble to implement multiple routes to the end of an analysis in most cases, so library() is usually best for me.
--
Sent from my phone. Please excuse my brevity.

On November 17, 2016 1:49:18 PM PST, Jeff Newmiller <[hidden email]> wrote:

>> require(tfplot)
>> tfplot(x.ts)
>
>Would just like to point out that require() should not be treated as
>interchangeable with library(). The former returns a logical status
>indicating success or failure, while the latter throws an error if it
>falls.  You should reserve use of require() for cases when you are
>implementing an alternative path of execution for failure, and in
>nearly all usual cases use the library() function instead so hapless
>users of your script don't have to sift through all the subsequent
>errors to figure out the problem.

______________________________________________
[hidden email] mailing list -- To UNSUBSCRIBE and more, see
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.