Re: Running Simulations and Demo Databases

Aaran (and others who undoubtedly have this same question):

We have not yet finalized a decision on whether to spend some time
converting the Eugene-Springfield 1980 database we used as a demo
database in UrbanSim3, or work out the data confidentiality issues with
using a more current database from one of the smaller current
applications of UrbanSim4, or to spend the time and effort to create a
made-up 'Smallville' dataset that would serve the purpose of running the
system. I'm leaning a bit towards the 'Smallville' approach but it will
take a bit of time and effort. Any thoughts on this from the user
community would be appreciated.

Paul

Aaran Stent wrote:
> David,
>
> Is there a working updated UrbanSIm 4 sample database? I assume that
> you would have used one for the workshop students? If we could get
> UrbanSim 4 simulations to run as easily as we did with UrbanSim 3 then
> the remaining issues may be easier to solve. A sample base year data
> set seems to be the missing piece of the puzzle.
>
> Regards,
>
> Aaran Stent
> National Economics
> aaran.stent_at_nieir.com.au <mailto:aaran.stent_at_nieir.com.au>
> 416 Queens Parade
> Clifton Hill VIC 3068
> Ph: 0419 227 262
>
> -----Original Message-----
> *From:* users-bounces_at_urbansim.org
> [mailto:users-bounces_at_urbansim.org]*On Behalf Of *David Socha
> *Sent:* Saturday, 15 July 2006 1:23 AM
> *To:* Joseph Nunez; users_at_urbansim.org
> *Subject:* RE: Running Simulations and Demo Databases
>
> Joe,
>
> The database section of the manual does not yet conform to the
> UrbanSim 4 standards. It is very close, but different in several
> ways.
>
> One of these ways is models_variables - this table is no longer
> used in UrbanSim 4. UrbanSim 4's integrated estimation process
> uses a Python file (for instance, see
> opus.psrc.estimation.estimation_HLCM_variables.py) to specify the
> set of model variables to use. The estimation process then
> creates coefficients and specification objects, which may be
> stored as tables in the database, so that a later simulation can
> re-create the coefficient and specification objects.
>
> We plan to add a tutorial on estimation, hopefully within the next
> few weeks.
>
> David
>
> David Socha
> Center for Urban Simulation and Policy Analysis
> University of Washington
> www.urbansim.org <http://www.urbansim.org/>
> 206 616-4495
>
> ------------------------------------------------------------------------
> *From:* users-bounces_at_urbansim.org
> [mailto:users-bounces_at_urbansim.org] *On Behalf Of *Joseph Nunez
> *Sent:* Thursday, July 13, 2006 9:32 PM
> *To:* users_at_urbansim.org
> *Subject:* Re: Running Simulations and Demo Databases
>
> Liming,
>
> Thanks for your help, you have certainly pointed me in the
> right direction.
>
> Reading through the user manual, the LPM model requires that
> its variables must be present in the model_variables. I am
> using an old model_variables table with the JAVA CLASS column
> still pointing to java paths not to OPUS. I have tried
> tinkering with this table changing path names and column names
> but to no avail, i will continue reading through the code to
> better understand how everything is pased from the db.
>
> That being said, Does anyone have the table schema
> for URBANSIM4 model_variables table??, ( in the user guide it
> merely refers to it).
>
> if not, I will patiently await for documentation of the
> Database tables section (Chapter 8) to be completed.
>
> Until then.....Hope you guys had a successful workshop.
>
> Best Regards,
>
> Joe
>
>
>
>
>
>
>
>
>
>
>
> -----Original Message-----
> From: "Liming Wang" <lmwang_at_u.washington.edu>
> To: "Joseph Nunez" <joseph.nunez_at_nieir.com.au>
> Cc: users_at_urbansim.org
> Date: Thu, 13 Jul 2006 14:02:20 -0700
> Subject: Re: Running Simulations and Demo Databases
>
> Joe,
> The error with land price model is because of a change of
> variable name format in model specification. You can refer
> to the user guide sections of variable name and
> specification and coefficients for the new format. We'll
> also provide some documents on converting from UrbanSim3
> to UrbanSim4 soon. For OpenEV you'll need to prepare a
> prototype_dataset and template OpenEV project file
> beforehand. It works for the application of PSRC in our
> environment with some file locations and names hard-coded.
> Because of compatibility issue, we may eventually phase it
> out and it won't be supported.
>
> Hope these help.
>
> Liming.
>
> On 7/12/06, *Joseph Nunez* < joseph.nunez_at_nieir.com.au
> <javascript:openWin('/WorldClient.dll?Session=GTWFZJC&View=Compose&To=joseph.nunez_at_nieir.com.au&New=Yes','Compose',800,600,'yes');>>
> wrote:
>
> Hi all,
>
> I have been exploring Urbansim 4. I have just
> encountered problems with running a simulation and
> creating indicators for the Eugene_Springfield
> baseyear 1980. I think this is attributed to that it
> was made for Ubransim 3 as the Input Database. Is
> there a sample database for Urbansim 4 (PSRC perhaps)?
>
> 1. If there isn't, I have started simple with just
> simulating the "land price model". I started a
> simulation but the following error: "No variable
> ('gridcell.devtype_1'). Looking through the
> opus.urbansim.gridcell , I have found only devtype_DDD
> . I do not know how/where this variable is computed
> from. (I assume devtype is Development Type)
>
> 2. In creating an Openev_map, do you need to save a
> opf file beforehand with all the shapefiles loaded
> within it? I recieve the following errors: unable to
> load 'template_region_uga_leg.opf' and get an empty
> tif file. I have tried opening the Shapefiles for the
> Eugene_Springfield in Openev but it just crashes the
> program. (I am on WinXP using Openev_FW)
>
> Any advice and comments are greatly appreciated.
>
> best regrads,
>
> Joe
>
>
>
>
> ---------------------------------------
> Running Land Price Model (from
> opus.urbansim.models.corrected_land_p
>
> rice_model): started on Thu Jul 13 12:57:22 2006
>
>
> opus.urbansim.gridcell.is_in_development_type_group_SSS
>
> Creating DevelopmentGroupSet object.
>
>
> opus.urbansim.gridcell.is_in_development_type_group_SSS:
> complet
>
> ed...0.0 sec, total 14740 KB, up 276 KB
>
> Total number of individuals: 15441
>
> LPM chunk 1 out of 1.: started on Thu Jul
> 13 12:57:23 2006
>
> Compute variables ...
>
> WARNING: No variable
> ('gridcell.devtype_1', 'devtype_1') fou
>
> nd or an error occured during loading.
>
> Traceback (most recent call last):
>
> File
> "C:\Python23\lib\site-packages\opus\urbansim\model_coordinators\model_sys
> tem.py", line 417, in ?
>
> s.run(resources, options.is_run_subset)
>
> ---------------------------------------------
>
>
>
> _______________________________________________
> Users mailing list
> Users_at_urbansim.org
> <javascript:openWin('/WorldClient.dll?Session=GTWFZJC&View=Compose&To=Users_at_urbansim.org&New=Yes','Compose',800,600,'yes');>
> http://www.urbansim.org/mailman/listinfo/users
>
>
>
> ------------------------------------------------------------------------
>
> _______________________________________________
> Users mailing list
> Users_at_urbansim.org
> http://www.urbansim.org/mailman/listinfo/users
>
Received on Mon Jul 17 2006 - 12:30:36 PDT