Bookmark and Share

Notice: On March 31, it was announced that Statalist is moving from an email list to a forum. The old list will shut down on April 23, and its replacement, statalist.org is already up and running.


[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: st: convert yearly data into monthly and interpolate


From   Nick Cox <njcoxstata@gmail.com>
To   statalist@hsphsun2.harvard.edu
Subject   Re: st: convert yearly data into monthly and interpolate
Date   Tue, 19 Apr 2011 11:50:41 +0100

Interpolation also implies filling in of missing values in gaps. Here
all your values are missing, as you have _no_ monthly values. I am
shocked that you don't appear to realise how difficult this problem
is, impossible even, and that you seem to be imagining that there is a
quick Stata fix that someone has thought up.

Whether you can or should ignore seasonality is an issue for your peers.

So my main point remains. There are various unsatisfactory ways to do
this and you haven't said how you want to do it (let alone why that
choice is appropriate).

There are any number of snake oil methods that might be thought up.
For example, an annual average might be assigned to mid-year and
monthly values then interpolated. Or, it might be assumed that
within-year variation is related to some proxy and monthly values then
assigned with the constraint that their averages by years match those
observed. See -denton- from SSC.

In some ways it would be simpler to -expand- each yearly average and
experiment with that series, and possibly various smooths of it.

Nick

On Tue, Apr 19, 2011 at 11:31 AM, D-Ta <altruist81@gmx.de> wrote:
> I would still call it interpolation (to me, imputation is related to the
> filling of missing values for some units only). But I guess my point is
> clear anyway.
>
> Moreover, seasonality patterns are not of importance for my approach, I
> really just want to know how to convert it to monthly format and interpolate
> over all month - preferably in a efficient way (!)
>
>
>
> Am 19.04.2011 10:23, schrieb Nick Cox:
>>
>> This is reminiscent of a question asked recently by Ben Ammar. See the
>> thread starting with
>>
>> http://www.stata.com/statalist/archive/2011-04/msg00131.html
>>
>> Perhaps terminology, or at least its use, varies between fields, but
>> your problem strikes me as one of imputation, not interpolation.  Any
>> way, it appears that there are various different ways to fudge this.
>> Which do you want? How do you propose to re-insert any seasonality
>> that has been lost?
>>
>> Nick
>>
>> On Tue, Apr 19, 2011 at 8:55 AM, D-Ta<altruist81@gmx.de>  wrote:
>>>
>>> Dear List Users,
>>>
>>> I have the following yearly data.
>>>
>>> id      Label   alq2_2000       alq2_2001       alq2_2002
>>> 1       group1  19,3            19,8            18,2
>>> 2       group2  19,3            19,8            18,2
>>> 3       group3  30,1            24,5            22,9
>>>
>>>
>>> the alq2_year variables measure group specific unemployment in a given
>>> year.
>>> I would like to merge it with another data set which consists of monthly
>>> data. Before converting my yearly data into a monthly format, I would
>>> like
>>> to interpolate it such that i get monthly interpolated estimates of alq2.
>>> i
>>> would appreciate if someone could help.
>>>

*
*   For searches and help try:
*   http://www.stata.com/help.cgi?search
*   http://www.stata.com/support/statalist/faq
*   http://www.ats.ucla.edu/stat/stata/


© Copyright 1996–2014 StataCorp LP   |   Terms of use   |   Privacy   |   Contact us   |   Site index