Open hanase opened 6 years ago
I wonder if it makes sense to have separate versions of datasources.py & models.py specifically for estimation and simulation. The reason being that datasources.py imports all the data, which will be different depending on whether we are running in estimation or simulation mode. I think this would be analogous to having two different configuration xmls in Urbansim 1.
Peter is going to write up a description of how the lag variables are created, so we'll have a better understanding of how the estimation data set differs from one used for simulation.
Thanks Stefan! I wonder how much code duplication this would require. But yes, for the estimation we need to figure out how to use lag variables (in both, simulation and estimation) before doing a big surgery.
@stefancoe - the estimation of REPM (and probably other models) is not working due to missing datasets. It's because these datasets are in the 2000 cache and not in 2014. I suggest to consolidate the estimation cache into two years:
I can work on it.
Figure out how a variable or model defined for a dataset, say persons, can be used on a dataset of different name, say persons_for_estimation.
Maybe via injecting that alternate dataset under the original name. Or, joining the two datasets?