I have always found it cumbersome to generate a table-like overview over group-wise descriptive stats out of panel/structured data in EVIEWS.
The program below generates a workfile with two variables x and y in a panel design in the page "panel" and a corresponding pool structure in the page "pool".
How can I quickly obtain a tabular presentation of some descriptive statistics, e.g., the mean, of the expression x/y for each of the cross sections, and perhaps also over two different samples, for example the samples s1 and s2 - either in the panel or in the pool structure?
Obviously, a longer program can do this, but could such a basic and often-needed task of data examination also be achieved faster with one to two command lines? Or is there an add-on around which would do the job?
(So, I am missing kind of the powerful and user-friendly "tabulate" command and by-group statistics as implemented in a well-known competitor of EVIEWS.)
Code: Select all
wfcreate(page=panel) q 2000q1 2015q4 10
pagestruct cid @date(dateid)
sample s1 "2000Q1 2008Q4"
sample s2 "2009Q1 2015Q4"
pageunstack(page=pool, namepat="?*") cid dateid @ x y