Lehman Center for Transportation Research at Florida International University
Research at the Lehman Center


A Transit Ridership Model Based on geographically Weighted Regression and Service Quality Variables

Fang Zhao, Lee-Fang Chow, Min-Tang Li, and Xuemei Liu
    This research is designed to investigate the spatial variations in the relationship between transit use and potential ridership predictors including demographics, socioeconomics, land use, accessibility, transit supplies, etc., in Broward County, Florida via GWR models. The transit use data were from the 2000 CTPP, which provided information about the percentage of workers that used transit mode to commute to work. The models indicate that regional accessibility to employment and percentage of zero-car households are global predictors, meaning that their impact on ridership is more or less uniform across the space. However, three more variables, employment density, average number of cars in households with children, and percentage of Black population, have various degree of influence over transit use as indicated by their coefficients and t-statistics. GWR models provide significant improvements in the accuracy of the prediction of transit use for HBW purpose over OLS regression models. Moreover, GWR allows areas to be identified where a particular variable has no predictive power of transit use. It is found that there exist different sets of predictors for transit use in different areas. However, the research also found that GWR models are not spatially transferable, therefore they cannot be applied to other urban areas directly because the spatial structure of an urban area, such as income distribution by location, also affects the location where a particular predictor may be significant as well the coefficients. Therefore, it is necessary to calibrate GWR models for each urban area. Using GWR models, transit use predictors may be identified for different areas, which may then be incorporated into ridership models. GWR also has the potential to improve ridership prediction at bus stop levels. Implementation of the methodology for this purpose is uncomplicated and does not require additional data.

Refinement of FSUTMS Trip Distribution Methodology

Fang Zhao, Lee-Fang Chow, Min-Tang Li, Albert Gan
    In this report, alternative trip distribution models are investigated with the purpose of improving the current trip distribution methodology in FSUTMS. Three types of models are studied: intervening opportunity models, enhanced gravity models, and destination choice models. The performance of the models was compared against travel survey data and that of traditional gravity models. The evaluation criteria included average trip length, trip length distribution, intrazonal trips, and spatial accuracy measured by different statistical tests. The findings from the project indicate that intervening opportunity models and enhanced gravity models did not produce noteworthy improvements, although the intervening opportunity models were handicapped by the lack of suitable software for model calibration therefore no definite conclusions may be drawn regarding their potentials. Aggregate destination choice models were developed for Broward, Palm Beach, and Volusia counties for Home-Based Work (HBW) trip purpose. Although the three counties are different in their demographics, socioeconomics, and urban structures, improvements are seen in the trip distribution results produced by all the destination choice models, although the improvements for Broward and Volusia counties are more significant than those for Palm Beach County. Results from destination choice models for other trips purposes also indicate improvements.

Alternatives for Estimating Seasonal Factors on Rural and Urban Roads in Florida

Fang Zhao, Min-Tang Li, Lee-Fang Chow
    The goal of this study was to develop a more objective and data-oriented approach by seeking to explain the underlying causes of seasonal fluctuation patterns in traffic data and to develop a methodology rooted in statistics for assigning seasonal factors to short period counts. Various clustering methods from the nonparametric hierarchical cluster analysis and parametric model-based analysis were evaluated in this study and those that better described the truth-in-data in the seasonal factor grouping process were identified. Conventional multiple linear regression analyses were conducted to identify possible factors contributing to the seasonal fluctuations in traffic volumes. Based on these factors, a fuzzy decision tree was developed to determine the seasonal factor group assignment for short period counts in the southeast Florida urban area. The methodologies developed in this study for seasonal factor grouping and assignment were implemented in a prototype GIS program, which also supported visualization of data on transportation system, traffic data, land use, and statistics about seasonal factors and groupings.

Optimization of Transit Network to Minimize Transfers

Fang Zhao, Albert Gan
    December 2003
    This report presents a mathematical methodology for transit route network optimization. The goal was to provide an effective computational tool for the optimization of a large-scale transit route network. The objectives were to minimize transfers and optimize route directness while maximizing service coverage. The formulation of the method consisted of three parts: representation of transit route network solution search spaces, representation of transit route and network constraints, and a search schemes capable of finding the expected global optimal result. The search methods included greedy search, fast hill climbing, and an integrated simulated annealing, tabu and greedy search algorithm. The methodology was implemented in a computer program, and was tested with several benchmark problems. It was also applied to a large-scale realistic network optimization problem in Miami-Dade County, Florida. The test results showed that the methodology developed in this study was capable of improving a given network solution in terms of average transfers and transit service coverage. A prototype application for transit network optimization was implemented with TransCAD as the user interface.

Development of FSUTMS Life Cycle and Seasonal Resident Trip Production Models for Florida Urban Areas

Fang Zhao, Lee-Fang Chow, Min-Tang Li, Albert Gan
    March 2003
    This study presents the procedure and results of analyses of household travel survey data from three participating MPOs/MSA: Lee County MPO, Volusia County MPO, and the Jacksonville MSA. The analyses involved calibrating trip production rates based on the structures of Tampa Bay lifestyle model, Southeast Florida lifestyle model, and the current standard FSUTMS model for home-base work (HBW) trip, home-based shopping (HBS), homebased social-recreational (HBSCR), and home-based other (HBO) trip purposes. The performance of the lifestyle models was evaluated by comparing the predictions from lifestyle models and those from the standard FSUTMS model against the expected trips based on survey and census data. Trip rates for seasonal households were also analyzed to determine if seasonal household shared similar travel behavior as retired households and if a separate set of trip rates were needed for seasonal households. The results from this study indicated that lifestyle models improved the trip production estimations for the four trip purposes for all three Florida urban regions to different degrees.

FSUTMS Mode Choice Modeling: Factors Affecting Transit Use and Access

Fang Zhao, Min-Tang Li, Lee-Fang Chow, Albert Gan, L. David Shen
    July 2002
    This study utilized the transit onboard survey data from the 2000 Southeast Florida Travel Characteristic Study to investigate the needs to incorporate additional variables other than travel time and costs into the modal split model of the Florida Standard Urban Transportation Model Structure (FSUTMS) to improve its transit forecasting capability. Factors including land use, demographics, the socioeconomic characteristics of the population, and transit supply were studied using regression analysis. GIS was used extensively for both data compilation and spatial analysis. The study identified a number of transit supply variables as well as population, employment density, and land use mix as strong indicators of transit use. Analysis of transit walk accessibility indicated that transit use beyond one half mile was minimum. A methodology was developed for estimating the percentage of transit service population in a zone, which proved to be a better indicator of transit use than the commonly used buffer method. Regression models that may be used for estimating and forecasting, for a future model year, the percentage of population and workers served by transit were also developed.

Generalized Method for Analysis of Pavement Management Database

Min-Tang Li, Lee-Fang Chow, and Irtishad U. Ahmad
    May 2002
    In Florida, pavement materials are periodically and systematically sampled and tested. The test results are recorded in the databases of pavement management system (PMS). The PMS databases are large and are currently stored on a remote mainframe system that is usually not directly accessible by an end user at a PC terminal with a modem connection. This project involves the development of a mainframe computer tool to allow the remote users to conveniently and instantly perform initial on-line statistical analyses on the FDOT's PMS databases. The SAS/AF software was used to create user-friendly interactive windowing applications. As a product of the Statistical Analysis System (SAS) software, SAS/AF provides interactive user interfaces to all the data access, management, analysis, and presentation features. The software uses the Frame class as the foundation for GUI-based applications. The Frame class provides windowing capabilities to SAS/AF applications. It can contain visual controls and non-visual components for creating a user interface. Through the Frame class, windows and dialog boxes, menu bars or banners, and application help can be created. However, SAS/AF versions 7 and up do not support build-time development of FRAME entries in the mainframe operating environments. Consequently, applications developed in other desktop environments must be ported to a mainframe, i.e., frames have to be created on a personal computer (PC) or other systems that support frame entries and then ported to the mainframe. The mainframe SAS/AF application program was designed to query a subset of records for analysis. This report documents the development process and serves as the user manual for the SAS/AF mainframe program developed. The procedure for porting the SAS/AF code from a PC to the mainframe is also included.

Technical Memorandum No. 3 for Refinement of FSUTMS Trip Distribution Methodology: "Calibration of an Intervening Opportunity Model For Palm Beach County"

Fang Zhao, Lee-Fang Chow, Min-Tang Li, Albert Gan, and L. David Shen
    September, 2001
    This technical memorandum documents an effort to calibrate an intervening opportunity model for Palm Beach County. Intervening opportunity models are investigated as an alternative to some of the gravity models currently used by some counties as a part of FSUTMS. This study has been motivated mainly by two reasons. First, Palm Beach County, like some other counties in Florida, has a narrow and long shape lying along the coast, with multiple city cores separated spatially. Gravity models tend to over-predict short trips in such cases. When the traditional gravity models do not perform well, K factors sometimes have to be used. Intervening opportunity models are believed to be able to better handle such urban development patterns. Another advantage of an intervening opportunity model over a gravity model is that it does not depend on a trip length frequency distribution as a gravity model does. For a gravity model, this trip length frequency distribution, typically obtained from a base year household survey, is assumed to remain unchanged for future model updates and forecasts. This assumption will be unlikely to hold in reality if there will be significant changes in future land uses. An intervening opportunity model uses attractions as the basis in model calibration, which may be updated when land uses have changed, thus making the model more adaptive to changes in trip patterns.

Integrating Land Use and Transportation in a GIS Visualization Tool

Fang Zhao, Min-Tang Li, Jill Strube, and Francisco Ordaz
    July 2001
    This report describes an effort to expand the prototype VOLUTI program into an integrated GIS-based tool that includes more land use and accessibility measurements, and additional functions related to assessment of impacts of land use developments and transportation projects. The improvements involved incorporation of additional data sources, development of land use and accessibility indicators, development of land use scenarios, and a stronger linkage between VOLUTI and FSUTMS. The tool incorporates a variety of databases, multimedia imaging, travel demand models, and useful evaluation methods to support visualization of land use and transportation information, and evaluation of land use and transportation interaction.

At-Grade Busway Planning Guide

L. David Shen, Hesham Elbadrawi, Fang Zhao, and Diana I. Ospina
    December 1998
    "This report reviews planning procedure for selected busway systems in North and South America, Europe, and other developing countries. Design issues to assure a safer operation of at-grade busway systems are also presented in this report. the information presented in this report in general and it can be modified according to the needs of each transit agency."

Analyses of Technologies and Methodologies Adopted by US Transit Agencies to Enhance Transit Security

L. David Shen, Diana I. Ospina, Fang Zhao, Hesham Elbadrawi
    December 1997
    "This research investigated successful security programs that have demonstrated to be effective in controlling and preventing transit crime, and enhancing passenger perception of security…The analysis of the researched data showed that the highest number of crimes occur against transit properties, followed by the crimes that affect security perception and the crimes against passengers respectively."

At-Grade Busway Study

L. David Shen, Diana I. Ospina, Fang Zhao, Hesham Elbadrawi
    December 1997
    "This research investigates the issues related to the implementation and operations of the South Dade Busway, the first at-grade busway in the United States. Busway applications in North America and through the world are also analyzed and presented through an extensive literature review. A case study about the South Dade Busway is also presented in this report."

Data Requirements and Data Sources for Transit GIS Applications

Fang Zhao, L. David Shen
    February 1998
    "This research involves the identification of the data needs for various transit GIS applications and possible sources from which useful GIS data may be obtained. Possible GIS transit applications, the data needs for these applications, the existing and potential data sources, and other issues related to data quality and maintenance are discussed."

Factors Influencing Future Transit Efficiency: Automated People Mover Cost Study

L. David Shen, Fang Zhao, Jian Huang
    December 1994
    "In this report, capital costs of five urban APM systems are analyzed and compared. Cost comparisons are also made between the urban and airport APM systems…The cost issue is also examined from the performance and planing perspectives. Cost reduction recommendations are made based on the results of the study."

Factors Influencing Future Transit Efficiency: Demographic Impact on Urban Guideway Transit Systems

L. David Shen, Yihua Xiong, Young-Kyun Lee
    December 1994
    "This report identifies and studies the demographic factors. To study demographic trends, data is gathered from the metropolitan areas of Miami, FL, Atlanta, GA, Greater Washington, D.C., and Boston, MA. The relationship between the guideway transit ridership and demographic characteristics is modeled using both traditional methods including correlation statistics and multiple regression analyses, and non-traditional methods including geographic information systems (GIS) and artificial neural network (ANN)."

Factors Influencing Future Transit Efficiency: Evaluation of the Demographic Impact on Guideway Transit Systems Using GIS and ANN Technologies

L. David Shen, Jian Huang, Young-Kyun Lee, Fang Zhao
    December 1995
    "In this report, the demographic impacts on urban guideway transit systems are studied. The relationship between demographic characteristics and guideway transit ridership is analyzed and visualized with demographic information system (GIS) technology. The relationship is also modeled using artificial neural network technology."

GIS Applications for Public Transportation Management

Fang Zhao, L. David Shen
    June 1997
    "In this report, results of a study on temporal GIS applications to various transit management tasks are summarized. The state-of-the-art of temporal GIS research and development is reviewed. Potential benefits of applying temporal GIS technologies to transit are investigated. A simple method to record the changes in spatial objects and their attributes with respect to time has been developed using PC ARC/Info relational databases and a means to query the spatiotemporal data."

Levels of service and Their Relationship with Rail Transit Capacity

L. David Shen, Young-Kyun Lee, Fang Zhao
    December 1995
    "In this report, the results of a study on the concept of LOS and the relationship between LOS and rail rapid transit capacity are presented. The study involves the examination of the service standards established by operators and transit operations practices. A LOS index model is also developed, which establishes LOS for rail rapid transit based on the vehicle load factor, headway, and travel speed."

Planning and Implementation of Automatic Vehicle Location Systems for Public Transit

Fang Zhao, L. David Shen
    June 1997
    "The current state-of-the-art AVL technologies, applications, and potential benefits are described in this report. Results from a survey conducted among 135 transit properties are summarized. The survey results show that smaller agencies are less likely to implement AVL since there are less incentives as congestion and crime problems are far less serious tan in larger urbanized areas."

Workshop on Emerging Fare Collection Technologies in Mass Transit

L. David Shen, Nii Attoh-Okine
    May 1997
    "The proceedings contains papers presented at the Workshop on Emerging Fare Collection Technologies in Mass Transit held in Miami, FL, September 26-27, 1996. The objective of the workshop is to provide a forum for the exchange of ideas and development in emerging fare collection techniques in Mass Transit."