Notice: On April 23, 2014, Statalist moved from an email list to a forum, based at statalist.org.
From | Nick Cox <njcoxstata@gmail.com> |
To | "statalist@hsphsun2.harvard.edu" <statalist@hsphsun2.harvard.edu> |
Subject | Re: st: Start/End dates to ongoing count indicators |
Date | Tue, 23 Apr 2013 17:18:19 +0100 |
I assume numeric variables with formats showing Stata weekly dates. First you need an observation identifier gen long obsid = _n and then a duration count gen duration = EndDt - StDt + 1 and then you want one observation per week expand duration Tidy up the dates bysort obsid : gen Week = StDt[1] + _n - 1 Then contract Week Class and finally it's a -reshape-. Nick njcoxstata@gmail.com On 23 April 2013 17:04, Jeremy Wells <jwell33@tigers.lsu.edu> wrote: > I have a dataset where observations are unique projects. Each project > has a start week and end week variable, as well as a classification variable, > so the data structure looks like this: > > StDt EndDt Class > 2002wk1 2002 wk3 1 > 2002wk4 2002wk6 1 > 2002wk1 2002wk5 2 > 2002wk4 2002wk5 2 > > What I would like to have is time series data with each observation a week > and variables for each class with a count of the number of ongoing projects > in each class in each week, so the data would look like this: > > Week Class1 Class2 > 2002 wk1 1 1 > 2002wk2 1 1 > 2002wk3 1 1 > 2002wk4 1 2 > 2002wk5 1 2 > 2002wk6 1 0 > > The first though is -tab class, gen(class)-, but that will give me a 1 for the > start week only, but I want the count to continue to the end week. * * For searches and help try: * http://www.stata.com/help.cgi?search * http://www.stata.com/support/faqs/resources/statalist-faq/ * http://www.ats.ucla.edu/stat/stata/