Difference between revisions of "Frequently Asked Questions"

From NFTPO Model
Jump to: navigation, search
 
(4 intermediate revisions by 2 users not shown)
Line 77: Line 77:
  
 
The Shadow Price step should be run when changes are made to the employment and/school enrollment data.
 
The Shadow Price step should be run when changes are made to the employment and/school enrollment data.
 +
 +
'''I noticed that when I rerun the model without any change to model inputs and scripts, the output trip tables are different each time. I would like to know if there is any randomness within the NERPM process, especially within the DaySim model?'''
 +
 +
When run using the exact same set of inputs, DaySim will produce the exact same set of outputs.  DaySim has a random seed setting that is used to seed the drawing of random numbers throughout the simulation.  Assuming this is the same, then DaySim will choose the same series of random numbers for each HH, person and choice model.  Sometimes users see differences because the work and school destination choice shadow prices are updated between runs, which changes the relative attractiveness of destinations. 
 +
 +
'''When I added some roadway improvement, some shifting in work location that were not quite related to the roadway improvement was observed. The number of trips destined for certain TAZs increased drastically. Is the work location choice process constrained by the number of employments in each parcel?  If one parcel has an employment of 100, is it possible that more than 100 workers choose that parcel as their work location?'''
 +
 +
The work and school location choice models are shadow priced, which means the chosen work and school locations are compared to the employment and enrollment by zone, a difference ratio is calculated, an additional shadow price (or cost) is added to each over-assigned zone, and then the process repeated until converged.  Destination choice shadow pricing is like auto-calibrating workers to jobs (and students to enrollment) by zone.  The relationship is not a hard constraint, i.e. it is possible for the workers that choose a zone to exceed the applicable employment.  It’s a good idea to instead look at a zone’s regional share of the relevant worker types to the relevant employment. 
 
   
 
   
 
'''How do you save the turning movements at a series of nodes?'''   
 
'''How do you save the turning movements at a series of nodes?'''   
Line 96: Line 104:
 
'''What is the ArcGIS Versions compatibility?'''
 
'''What is the ArcGIS Versions compatibility?'''
  
The Generate of Edit Parcel File Application developed for use with the NERPM-AB model has been tested and is compatible with ArcGIS 10.2.  The Generate of Edit Parcel File Application might not be compatible with the newer versions of ArcGIS.  However, recently the tool was tested with ArcGIS 10.4, and with the following changes in the code, the tool will be compatible with the latest version.
+
The Generate of Edit Parcel File Application developed for use with the NERPM-AB model has been tested and is compatible with ArcGIS 10.2 .  The Generate of Edit Parcel File Application might not be compatible with the newer versions of ArcGIS.  However, recently the tool was tested with ArcGIS 10.4 ([http://northfloridatpo.com/images/uploads/wiki/prepare_workspace.py prepare_workspace.py])  and with the following changes in the code, the tool will be compatible with the latest version.
  
 
New code statements:
 
New code statements:
 +
 +
import csv
 +
    with open(output_parcels_csv,'wb') as f:
 +
      w=csv.writer(f)
 +
      w.writerow(fieldnames)
 +
      for row in arcpy.da.SearchCursor(output_parcels,field_names=fieldnames):
 +
        w.writerow(row)

Latest revision as of 16:50, 20 February 2019

When editing the parcel attributes (via the parcel editing tool), you can edit the number of households. Can you also/do you need to indicate if these households are single- or multi-family?

No, the user can only edit the number of households with the editing tool. The characteristics of the households are assigned by the tool. The trip making patterns of the household are based on the characteristics of the household and the persons within the household, not by identifying the type of dwelling unit.

Is there an attribute for hotel/motel population or number of rooms?

The NERPM-AB does not contain a visitor’s model and the trip making patterns are associated with the residents of the area. The only way to associate trips with a hotel/motel is by adding employment associated with the use. In developing the 2040 socioeconomic data, the ITE trip rate of 0.90 commercial employees per room was used. These rates change and can vary by type of hotel. The user should determine if this rate is appropriate for new hotel/motel units added. The Shadow Price step should be run when changes are made to the employment and/school enrollment data.

Note: The current version of the NERPM-AB, uses special generators to account for the number of trips associated with Jacksonville International Airport and the City of Saint Augustine.

Are there restrictions associated with the naming conventions?

Yes, avoid blank spaces in file name paths. Although hyphens will work, using underscores is a better way to create clarity in the naming conventions.

Will work:

WorkingDirectory="D:\FSUTMS\NERPM_AB1\Master\Base2010\Output\DaySim\working"

Will NOT work:

WorkingDirectory="D:\FSUTMS\NERPM AB-1\Master\Base2010\Output\DaySim\working"

What programs do I need to run PopGen?

We did not provide PopGen with the modeling package that has been made available through the FSUTMS portal because the user does not need to run PopGen. PopGen is not part of the distribution, and we do not recommend that the end users attempt to use it to produce new synthetic populations. It should only be run by RSG, NFTPO and D2 staff to produce baseline populations.

PyQt is a Python GUI library, which is used to create PopGen’s user interface. The version of PyQt to which the questioner refers to is an older version that is probably no longer supported. It may be possible to download and install a more recent version of the library, but we cannot guarantee that it would be backward compatible. In addition, it would be possible to run PopGen without a GUI, but that requires some reprogramming to break the dependencies on the UI and to run the program from command line inputs.

The current, active curator of PopGen is Prof. Karthik Konduri at the University of Connecticut (kkonduri@engr.uconn.edu).

All DaySim future year parcel and synthetic household and person files, covering the years 2011 through 2039, have been created and are available to the user. They are created in the form of Geodatabases which are part of the NERPMAB1v2 download. The Geodatabase files are located in the Input_ParcelData folder. The years 2010, 2030 and 2040 DaySim input files are included in the current version of NERPMAB1v2. Please refer to the section “Editing Tool” in the NERPMAB1v2 documentation or to the video link in the "Support" section for guidance on converting the Geodatabase files into the correct format for use by the DaySim program. These two sections also provide the user with directions on how to change the socioeconomic data files.

Does the NERPMAB1v2 run in Cube 6.4.1?

No, currently the NERPMAB1v2 is not compatible with Cube 6.4.1. Model runs are being made to analyze the issues. When resolved, we will update the answer to this question.

Can the data (_jax_hrec2040.dat) be summarized into SFDU and MFDU?

No, DaySim does not use dwelling units. Instead it uses households and "focusses" on the characteristics of the household; not the dwelling unit.

What future enhancements are planned for the model?

  1. Update of truck data
  2. Tolling for managed lanes
  3. Data collection-In FY 2016/17 and FY 2017/18 household travel data and origin-destination data for update of the model for the 2045 Long Range Transportation Plan which is due in Fall of 2019. This data will improve validation/calibration and estimation of EE, EI and IE trips.

What should I do if I find an incorrectly coded link and/or links with zero volumes?

Please report incorrect coding for correction in a subsequent model release. If you are doing a sub-area study and have links with zero volumes it may indicate a coding error, the need for a new centroid connector… Please let us know how you resolved the problem.

When setting up the new NERPMAB1v2 model is a full installation of ArcGIS is required?

Yes, you will need the full ArcGIS 10.2 or higher to run the land use and parcel editing tool. Please install ArcGIS 10.2 or higher and reinstall CUBE without ArcGIS.

ArcGIS Desktop software has 3 levels of licensing. Basic, Standard and Advanced. Which level of licensing is required for NERPMABMv2?

You need ArcGIS 10.2 or higher with basic or standard licenses but not Advanced.

The property appraiser updates the parcel data annually. How is the user supposed to keep the parcel data in the model current? Will the department provide annual releases of the parcel data corrected to the population and employment?

Property appraiser information is updated annually. However, the ownership is typically updated, and not the parcel boundary. The ABM allows for any year to be analyzed and the end user should verify the parcel boundary with the Property Appraiser (website) to ensure no boundary changes have not occurred. If no changes have occurred, then input the new land uses or trips. For future year analyses, the most current Parcel layer is not needed. Consult with the agency you are working with. The data sets are updated every five years in development of the long range transportation plan. We do not have the resources to update them annually

When you add a new development, in your example you are only adding development to one parcel. Won’t that just add development to an existing zone and lump the data of the parcel in with all the other parcels of that zone?

Yes, that is what happens. If you want to know difference between the background traffic and the traffic from the development, you run a select link analysis without the development and with the development to analyze the difference. It was suggested that a new TAZ be created with just the one parcel tied to it, for this analysis. We have not verified that the results are the same with both methodologies. We will follow up on the Wiki FAQ page. We suggest that you consult with the agency you are doing the analysis for to make sure they concur with the methodology you use.

I have a mixed use project with multiple TAZs and have tried to re-run the highway module to get distributions for each of the TAZs but get an error. Any suggestions to get multiple distributions for the same model scenario.

It was suggested that the Path File options could be used for this. You can then use the Cube analysis tools to get multiple select links from the same assignment.

We don't have ArcGIS so how do you change things manually?

The instructions to manually edit the files are included in the User’s Guide on the wiki page. Due to the interrelationship of the four DaySim files and the need to edit each file, we strongly recommend using the Parcel and Land Use editing tool.

When should I run the Shadow Price step?

The Shadow Price step should be run when changes are made to the employment and/school enrollment data.

I noticed that when I rerun the model without any change to model inputs and scripts, the output trip tables are different each time. I would like to know if there is any randomness within the NERPM process, especially within the DaySim model?

When run using the exact same set of inputs, DaySim will produce the exact same set of outputs. DaySim has a random seed setting that is used to seed the drawing of random numbers throughout the simulation. Assuming this is the same, then DaySim will choose the same series of random numbers for each HH, person and choice model. Sometimes users see differences because the work and school destination choice shadow prices are updated between runs, which changes the relative attractiveness of destinations.

When I added some roadway improvement, some shifting in work location that were not quite related to the roadway improvement was observed. The number of trips destined for certain TAZs increased drastically. Is the work location choice process constrained by the number of employments in each parcel? If one parcel has an employment of 100, is it possible that more than 100 workers choose that parcel as their work location?

The work and school location choice models are shadow priced, which means the chosen work and school locations are compared to the employment and enrollment by zone, a difference ratio is calculated, an additional shadow price (or cost) is added to each over-assigned zone, and then the process repeated until converged. Destination choice shadow pricing is like auto-calibrating workers to jobs (and students to enrollment) by zone. The relationship is not a hard constraint, i.e. it is possible for the workers that choose a zone to exceed the applicable employment. It’s a good idea to instead look at a zone’s regional share of the relevant worker types to the relevant employment.

How do you save the turning movements at a series of nodes?

Turning movements for specific nodes are saved by listing the node number in the assignment scripts (HTHWY00A.S, HTHWY00B.S, HTHWY00C.S, and HTHWY00D.S). The node numbers are hardcoded into the assignment scripts on the lines 26 and 27. In the current NERPMAB model, line 26 is ignored through the use of the semicolon. The semicolon turns the line into a comment, which is ignored by the program. Line 27 identifies the specific turns the user would like to save.

For example:

TURNS N= 1-999999

TURNS N=39782,40555,40265,40258,40241,40568

Should the user want to save all the turns, then remove the semicolon from line 26 and start line 27 with the semicolon. In this example, the turning movement for the listed node numbers (39782, etc.) will be saved.

After modifying the assignment scripts, save the scripts and run the model to obtain the turning movements. The turning movements will be save in the *.bin files and the volumes can be displayed by selecting the “intersections” tab in Cube and opening the desired *.bin file to display the volumes.

Can I still use the Trip Impact Tool to calculate the difference between the Build and No Build scenarios if there are differences in the highway networks?

Yes, as long as the links exist in both networks. The NERPM-AB network is a database network, so all nodes and links are the same in all networks. However, facility types that are coded with a zero, will not be carried over into the loaded networks. If you add a facility to the Build network and leave the facility type set to zero in the No Build network, the difference between the two networks cannot be calculated. One way around this, is by coding the facility type the same in both networks but setting the number of lanes to zero in the No Build network. This way there will not be any volume on the links with the zero lanes and the application will still work.

What is the ArcGIS Versions compatibility?

The Generate of Edit Parcel File Application developed for use with the NERPM-AB model has been tested and is compatible with ArcGIS 10.2 . The Generate of Edit Parcel File Application might not be compatible with the newer versions of ArcGIS. However, recently the tool was tested with ArcGIS 10.4 (prepare_workspace.py) and with the following changes in the code, the tool will be compatible with the latest version.

New code statements:

import csv
    with open(output_parcels_csv,'wb') as f:
      w=csv.writer(f)
      w.writerow(fieldnames)
      for row in arcpy.da.SearchCursor(output_parcels,field_names=fieldnames):
        w.writerow(row)