September 2025 Meeting

Natalia Brown · August 22, 2025

Meetings

Date/Time: Thursday September 11, 2025 10:00-11:30am
Location: MAG - 586 E 800 N, Orem, UT 84097

Virtual option available through the calendar appointment. Email utahMUG@gmail.com to request the calendar appointment.

Agenda

Welcome & Introductions

  • Anyone attending for the first time?
  • User Spotlights

Presentations

  • Leveraging mobility data for the automated calibration of travel demand models
    Gaurav Vyas, Bentley Systems.
    Summary: Travel demand models whether 4-step or Agent-Based (ABMs) were traditionally based on special travel surveys such as Household Travel Surveys (HTS’s) that served as the main source for estimation of the model coefficients. Other types of data such as transit on-board surveys, traffic counts, transit ridership by line, etc., were primarily used for the model validation and possible adjustments of the coefficients. The presentation illustrates a new a new general approach where different types of data including HTS and big data, can be used in one systematic process of travel model calibration in an automated manner using the case studies from MAG Weekend ABM and Chattanooga Person Travel Demand model.

  • Half Time Break
    Jared Lillywhite, Mountainland Association of Governments

  • Enhancing CRT Modeling in WFRC TDM v9.2: Diagnosing and Correcting Station-Level Forecast Biases (link)
    Bill Hereth, WFRC
    WF TDM v9.2 aims to improve CRT modeling, addressing major underestimation of Utah County boardings (-34%, Provo -90%) and overestimation in Davis County (+58%). Refinements include adjusting drive-access utility (RUNFACTOR = 2.5) and adding purpose/period ASCs by distance bins, which reduced CRT trip distance errors from -11%–-56% to -5%–-18%. While large in-vehicle time adjustments (60–75 minutes) aligned Provo boardings, they harmed model sensitivity. Findings will guide users on CRT limitations and support calibration with 2023 Household and Transit On-Board Survey data.

Discussion Topics

  • Utah-Specific resources:
    • Volume traffic map (link)
    • Household/Job viewer (link)
    • Model one-pagers and documentation (link)
  • Model Status Update (link)
    • Utah Statewide Travel Model: Hayden Atchley, UDOT
    • Wasatch Front TDM: Suzie Swim, WFRC
    • Cache TDM: Isaac Gardner, CMPO
    • Dixie TDM: Radhika Bhandari, DMPO
    • Summit-Wasatch TDM: Hayden Atchley, UDOT
    • Iron TDM: Hayden Atchley, UDOT
  • Presentation solicitations (link) – due by October 31st
  • Others?

Next meeting

  • Date/Time: Thursday, January 22, 10:00-11:30 am @ UDOT.

🍽 lunchiMUG

  • Gather with us at Kneaders for lunch.

Notes

Presentation - Leveraging mobility data for the automated calibration of travel demand models Gaurav Vyas, Bentley Systems

  • One Model Platform, Many Models
    • Assemble virtually any travel demand model structure including trip-based, tour-based, hybrid and activity-based models
    • Maintain different models or versions in parallel
    • Reduce time/effort to develop a new travel model
    • Adapt or upgrade models with advanced features over time
  • Harmonized Demand Modeling with OpenPaths
  • Common Data Sources in Model Development
    • Household travel survey data
    • Traffic counts
    • Transit ridership
    • Primary only used for model validation
  • Big Data as a replacement?
    • Pros
      • Becoming increasingly available from vendors
      • Big dta trip tables can be used to aggregate 4-step models in practice
    • Cons
      • Not behavioral
  • How is big data used for manual calibration in practice to far?
    • Pre-processing of O-D level data to create sub-model specific targets
    • No systemic approach to identify outliers
    • Our approach
      • Use o-d data directly for model calibration
  • Data Fusion
    • HTS
      • Total number of tours/trips/activites
      • Aggregate car ownership, mode share, TOD
    • Transit OB Survey
  • Calibration Instrumentation
    • Lots of different parameters that you can change for each element
    • How much should we change, and by how much?
  • Automated Calibration
    • Machine learning approach for better calibration results than manual calibration
    • Accelerates model calibration work
    • Enables data fusion from multiple sources
    • Modelers stay in control with transparency on the adjustments
    • Equally applicable to four-step and ABMs.
  • MAG (Arizona) Weekend Model
    • Calibration targets
      • Weekend activity rates
      • Big Data O-D tables
        • AirSage Data
      • Traffic counts
    • R Squared & %RMSE
      • Improved from .76 to .82 after 10 iterations
      • Not a huge improvement, but substantial improvement in comparison to doing it by hand
      • Final results - .73 to .92, RMSE from 54% to 28%
    • Get calibrated coefficients
      • Shopping, maintenance, eat-out, visiting, discretionary
  • Chattanooga
    • Calibration via patterns
    • Compromise between patterns & travel survey
  • Outcomes
    • All travel model simulation results are consistent with the data used in the automated calibration process
    • All data sets are validated against a common structure
  • Coming Soon: Traffic Assignment
    • 11x speedup vs Cube Voyager Highway
    • Results proportionality, consistency
    • New Analysis
      • Integrated ABM
      • Smaller file size
  • Calibration with Big Data
    • Don’t want to rely only on the big data, want to focus on a fusion of other datasets that you have.

Presentation - Enhancing CRT Modeling in WFRC TDM v9.2: Diagnosing and Correcting Station-Level Forecast Biases Bill Hereth, WFRC

  • Model Imprevement Investigations for Regional Rail (Formerly CRT)
  • Wasatch Front TDM Realities
    • Station level accuary is unrealistic, because transit is only calibrated and validated at the mode level
    • An improvement attempt in 8.3.2 was unsuccessful
    • With the upcoming activity based model (ABM) project, the model choice model will be re-estimated, offering another opportunity
  • Needs for Improvement
    • Boardings are low in Utah County and high in Davis County
      • Provo Station is 91% low
    • Drive to transit distance for regional rail in model is high for Utah County
      • Orem 85% high
    • Distance Traveled is Low
      • Model distances are low across the board
  • Key Terms
    • Run Factor – parameter used in transit path-building and assignment that adjusts perceived in-vehicle time by mode
    • Max Time – Drive to Transit – maximum allowable drive time from a traveler’s origin to a transit access point before the model assumes the person would not reasonable choose transit
    • Alternative Specific Constant (ASC)
      • Calibration parameter in discrete choice models that captures the average preference for an alternative not explained by observed attributes
        • Adjusts model predictions to better match observed mode shares
        • Accounts for unmeasured factors (comfort, image, reliability, etc)
    • Trip Length Frequency
  • Main conclusions
    • Adjusted RUNFACTOR to 2.5 for drive to transit and drive access MaxTime align median drive access distance for Utah County Stations
    • Purpose/period-specific alternative specific constants by distance bins significantly improve
    • If you manually make Provo much more attractive it causes deficiencies to the rest of the model
  • Overall results
    • The % variance is significantly better
  • Sensitivity Tests
  • Recommended Improvements
    • Adjust drive access utility for CRT
      • Runfactor =2.5
      • Better aligns modeled behavior with observed travel patterns, particularly at end of the line stations (Provo and Ogden)
    • Redefine regional rail distance on transit modeling
      • Apply alternative specific constants ASCs by distribution bins
      • Improves fit to observed distributions of trip distances on CRT
    • Document lessons learned for v9.2 calibration and future model application
  • Questions
    • Is it possible that we’re missing a subset of the population? Possible that it’s a bit of a mode-choice issue for that subset of the population

Model Status Update

  • Utah Statewide Travel Model: Hayden Atchley, UDOT
    • 2023 Base Year inputs
  • Wasatch Front TDM: Suzie Swim, WFRC
    • 2023 Base Year inputs
    • Microtransit
    • New RTP edits
    • Commercial trucks
    • V9.2 by the end of the year
  • Cache TDM: Isaac Gardner, CMPO
    • 2023 Base Year inputs
      • HH TAZ verification
  • Dixie TDM: Radhika Bhandari, DMPO
    • 2023 Base Year inputs
    • Mismatches in HH
    • Recreational factors
  • Summit-Wasatch TDM: Hayden Atchley, UDOT
    • 2023 Base Year inputs
    • Freight components
  • Iron TDM: Hayden Atchley, UDOT
    • 2023 Base Year inputs
      • Probably done next week
    • Freight component adjustments pending