ARIZONA DEPARTMENT OF TRANSPORTATION RESEARCH CENTER APRIL 2016 RESEARCH RESULTS DIGEST SPRͳ736: LONGͳDISTANCE TRAVEL MODELING: PROOF OF CONCEPT The Arizona Statewide Travel Demand Model (AZTDM) is one of ADOT’s key technical analysis tools for long-range transporta on planning. It is also used in the assessment of alterna ve mobility investments such as intercity transit and engineering design studies and traffic opera ons analysis. One purpose of the AZTDM is to es mate long-distance intercity trips between Arizona’s small and large urban areas. To achieve reliable es mates, the model must reflect actual travel behavior. These trips, longer than 50 miles, are es mated using data from the 2002 Na onal Household Travel Survey. However, this survey is aging and may not capture recent trends in travel behavior, and its sample of long-distance travel in the Southwest is small. For example, two of Arizona’s nearest neighbors, Nevada and New Mexico, have no NHTS long-distance travel records. The current model bases travel behavior between Arizona and these states on travel survey records synthesized from other neighboring states. Another element of long-distance personal travel is the extreme commute, defined as more than 50 miles in distance or more than one hour of travel. Extreme commutes between cities like Phoenix and Tucson cross regional boundaries. While census journey-to-work data shows that thousands commute between Phoenix and Tucson, the AZTDM home-based work model does not adequately predict this travel behavior. RESEARCH OBJECTIVES AND SCOPE This research project was established to provide ADOT with direc on on the best sources of data and best prac ces for upda ng its long-distance personal travel models to be er reflect observed travel behavior. Its original intent was to recommend approaches for combining these data sources ADOT RESEARCH CENTER 206 S. 17th Avenue Mail Drop 075R Phoenix, AZ 85007 PRINCIPAL INVESTIGATOR Lee Engineering 3610 N. 44 Street, Suite 100 Phoenix, Arizona 85018 to improve the model and ensure that long-distance trips are reflected in forecasts. The research scope comprised the following tasks: • Iden fy advances in long-distance personal travel modeling and feasible new sources of long-distance travel data by reviewing published literature; iden fying private data sources, such as AirSage and INRIX; and interviewing industry leaders. • Establish evalua on criteria, and assess the suitability and feasibility of the iden fied data sources for ADOT’s use in es ma ng long-distance travel. Iden fy the 2-3 highestranking data sources. The first phase of research concluded a er this task, and a decision was made to not proceed to the proof-of-concept tasks, as described below: Using a purchased sample from each of the recommended data sources, conduct a proof-of-concept test to determine the poten al of each source to improve ADOT’s ability to es mate long-distance travel. Based on the results of the proof-of-concept tes ng, develop recommenda ons regarding the best data sources and sta s cal techniques for es ma ng longdistance travel in Arizona. CONTACT US azdot.gov/research ResearchCenter@azdot.gov 602.712.4052 Table 1. Traditional Data Sources Data Source ACS data Features Demographic, socio-economic, and land-use data Demographic, socio-economic, and trip data Potential Reliable population data for both four-step and activity based modeling Purpose Generation of base population Evaluations Currently using Has potential to transfer modeling parameters Model development Highly recommended for transferability study NHTS 2016 Demographic, socio-economic, and trip data Traffic volume data Model development, calibration and validation Model validation Highly recommended BTS data HPMS data Traffic volume data Traffic volume data Traffic volume data Traffic volume data Origindestination and traffic volume data Demographic, socio-economic, and trip data Reliable, most recent and comprehensive data for both four-step and activity based modeling Air passengers and border crossing traffic volume Traffic volume for screen line validation Traffic volume for screen line validation Traffic volume for screen line validation Traffic volume for screen line validation Intercity transit passenger flow Model validation Model validation Model validation Model validation Model calibration and validation Currently using Ingress and egress models, and information on non-residents Model development, calibration and validation for air passengers Origindestination and traffic volume data Demographic, socio-economic, and trip data Traffic volume for screen line validation Model validation Highly recommended (if disaggregate level modeling technique is adopted) Currently using Supplement information on long distance personal travel Trip matrix Comprehensive trip matrix for the entire country Activity duration Model development, calibration and validation Model calibration and validation Not applicable CHTS data (long distance trip component) CCS data FMS data DPS data Intercity Passenger travel data Airport Ground Survey data Tourism Industry data Roadside intercept survey data FHWA traveler analysis framework data American time use data Trip information (to some extent) Currently using Currently using Currently using Currently using Highly recommended Recommended for pilot study Recommended for further study Not recommended Table 2. Emerging Data Sources Data Source Features Potential Purpose Evaluations AirSage data Trip information, and demographic and socioeconomic data Model calibration and validation Model calibration and validation Recommended for proof-ofconcept study INRIX and StreetLights data Trip information, network configuration, and demographic and socioeconomic data Network GIS file, and model calibration and validation Model calibration, validation, and GIS network file Recommended for proof-ofconcept study FDOT Smartphone Apps for ADOT Demographic, socioeconomic, and trip with exact location information Easy customization since it’s an open source program Model development, calibration and validation Recommended for ADOT’s adoption NHTS 2016 Survey with GPS sampling Demographic, socioeconomic, and trip with exact location information Model calibration and validation Model development, calibration and validation Costly option and has been excluded from the plan Bluetooth/Wi-Fi data Origin-destination, and traffic volume and speed Traffic volume for screen line and condor line validation Model calibration and validation Recommended for Pilot test bed ALPR data Origin-destination, and traffic volume and speed Traffic volume for screen line and condor line validation Model calibration and validation Recommended for Pilot test bed HERE data Traffic volume and speed, and transportation network configuration file Transportation network configuration Improvement of network configuration Recommended for network GIS file LBSN data Origin-destination, and traffic volume Not significant at this point Not applicable Not recommended Travel website data Origin-destination, and trip information Not significant at this point Not applicable Not recommended Delcan Technologies data Origin-destination, and traffic volume and speed Not significant at this point Not applicable Not recommended Cellint data Origin-destination, and traffic volume and speed Not significant at this point Not applicable Not recommended TomTom PND data Origin-destination, and traffic volume and speed Not significant at this point Not applicable Not recommended RFID data Origin-destination, and traffic volume and speed Not significant at this point Not applicable Not recommended RESEARCH RESULTS Tables 1 and 2 summarize the results of the research team’s inves ga on into poten al data sources for long-distance travel es ma on, which were acquired through a literature search and interviews with industry leaders. The tables also contain the team’s recommenda ons based on its evalua on of the suitability of the data for ADOT’s use. ACRONYMS USED IN TABLES 1 AND 2 ACS American Community Survey ALPR automated license plate recogni on BTS Bureau of Transporta on Sta s cs CCS con nuous coun ng sta ons CHTS California Household Travel Survey DPS Department of Public Safety (Arizona) FDOT Florida Department of Transporta on FHWA Federal Highway Administra on FMS freeway management system HERE product name; not an acronym HPMS Highway Performance Monitoring System INRIX product name; not an acronym LBSN loca on-based social networking NHTS Na onal Household Travel Survey PND personal naviga on device RFID radio frequency iden fica on PROOF OF CONCEPT The research team recommended the tes ng of data purchased from AirSage and INRIX if the study were to proceed to the proof-of-concept stage. The plan for the proof-of-concept was intended to determine whether using these data: • Eliminates the current limita ons of AZTDM • Changes the results of modeling using AZTDM • Increases the accuracy of the model and, if so, the significance of the impact • Is cost-effec ve for ADOT ADOT decided not to proceed with the proof-of-concept study using AirSage and INRIX data ini ally contemplated in the work plan developed for this study.