It’s implemented as an integer binary (i.e., 0 or 1) field labeled ‘for_estimation’ in the psrc_2014_urbansim2_estimation.jobs table. Jobs coded as a 1 correspond either to new locations since 2010 (in which case all workplace employment qualifies), or to jobs above 2010 levels at pre-existing workplaces. Public education and government are not located via ELCM so their lack of representation should not be problematic.
The 2010-2014 window is a larger one than our last jobs_for_estimation dataset, and corresponds directly to the baseyear jobs table (in contrast, the “growth timeframe” for our 2000-baseyear jobs_for_estimation dataset was only a year or two, and from a period later than the baseyear—2005-07, I think). In both cases the timeframe corresponds to a recovery from recession rather than “normal” economic expansion (if “normal” exists); hopefully there aren’t unique aspects as a result that are problematic when generalized over long periods, but this may be worth keeping in mind.
Email from Mike 01/25/2017:
It’s implemented as an integer binary (i.e., 0 or 1) field labeled ‘for_estimation’ in the psrc_2014_urbansim2_estimation.jobs table. Jobs coded as a 1 correspond either to new locations since 2010 (in which case all workplace employment qualifies), or to jobs above 2010 levels at pre-existing workplaces. Public education and government are not located via ELCM so their lack of representation should not be problematic.
The 2010-2014 window is a larger one than our last jobs_for_estimation dataset, and corresponds directly to the baseyear jobs table (in contrast, the “growth timeframe” for our 2000-baseyear jobs_for_estimation dataset was only a year or two, and from a period later than the baseyear—2005-07, I think). In both cases the timeframe corresponds to a recovery from recession rather than “normal” economic expansion (if “normal” exists); hopefully there aren’t unique aspects as a result that are problematic when generalized over long periods, but this may be worth keeping in mind.