Re: [Gretl-users] aggregate or disaggregate
On Tue, 9 Dec 2008, Hodgess, Erin wrote:
Are there any functions to either aggregate or disaggregate
data, please?
If so, would there be any interest in them, please?
There are no functions in gretl specifically geared to aggregation
or disaggregation of data, although
* there are functions that handle compaction and expansion of time
series data between specific frequencies, and
* there are various functions designed for handling panel data
that could be used for this purpose, and
* gretl's scripting capacity (noting in particular the possibility
of sub-sampling and looping) is very flexible.
More specifically, if one had disaggregated data (say at the level
of firms) and a variable that records the industry to which each
firm belongs, it would be easy to script an aggregation to
industry level using sub-sampling (the "smpl" command) and looping
(the "loop" command). One could (say) sum or average the
firm-level data to generate aggregate industry-level data, with or
without weighting.
As for disaggregation of data, I have to say that I'm not sure
what you mean. I'd have thought that to get disaggregated data
you'd have to start with disaggregated data, but maybe I'm missing
something.
If you have concrete ideas on these topics, please let us know,
and if they look useful we'll be happy to consider incorporating
them.
Allin Cottrell