2013 Session: 424

• Comparative Study of Alternative Methods for Generating Route-Level Mutually Exclusive Service Areas
  Abstract: This study investigates the willing-to-walk distance for measuring spatial accessibility of bus stops and examines the effectiveness of alternative methods for generating mutually exclusive transit service areas at the route level. We first investigate the walking distance to and from a transit stop using on-board survey data. Two methods in Geographical Information Systems (GIS), the Combination of Thiessen Polygon and Buffer (CTPB) and the Network Distance-based Service Area (NDSA), are compared as alternative strategies for generating mutually exclusive service areas (MESAs). To examine the effectiveness of these two methods, all MESAs are validated using a spider diagram generated from an on-board survey. Measures of urban form are also statistically tested for comparison of the two different methods. A case study of a single route, serving the Minneapolis-St. Paul metropolitan area, is performed using data from various sources, such as Google’s General Transit Feed Specifications (GTFS), on-board survey, parcel-level land uses, and street network. The validation with on-board survey data demonstrates the strengths of each method. The results also show that the NDSA, a popular GIS method for service area analysis, does not yield a more meaningful strategy for generating mutually exclusive transit catchment areas, especially when the spacing between stops is very small.
  Authors: Lee, Sanggu; Tong, Daoqin; Hickman, Mark D.
  Authors: Lee, Sanggu; Tong, Daoqin; Hickman, Mark D.
  Year: 2013
  Document Type: Paper
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-2803
• Riding More Frequently: Disaggregate Ridership Elasticity Estimation for a Large Urban Bus Transit Network
  Abstract: In this paper, stop level transit elasticities with respect to service frequency are estimated and discussed. Public transportation ridership is typically studied at an aggregate level, where variables influencing ridership are averaged over time and space for a metropolitan area. Understanding transit ridership at a finer temporal and spatial level is generally limited to mode choice models. Most aggregate analyses are unable to capture important effects at the parcel or block level. Such analyses also cannot account for variation in demand over time of day, an issue which has been addressed to some extent via time series modeling. Using data for the Chicago transit system, the results suggest that aggregate analyses overestimate the effect of service frequency on demand. In the context of other disaggregate analyses, these results suggest that walking quality results in distinct increases in ridership, even after accounting for land use, population and other demographics. A headway elasticity of ridership is estimated to be -0.263 to -0.277, very similar to recent disaggregate analysis of New York City transit data. The case is made for a better spatio-temporal understanding of transit ridership in order to allocate resources effectively.
  Authors: Frei, Charlotte; Mahmassani, Hani S.
  Authors: Frei, Charlotte; Mahmassani, Hani S.
  Year: 2013
  Document Type: Paper
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-3413
• Time-Varying Route-Level Transit Patronage Model
  Abstract: This study presents a time-varying transit patronage model, capturing the interaction between transit demand and land use patterns. This study considers how one might evaluate the transit demand changes by time of day with respect to land uses. Depending on a specific time of day, land uses (e.g., residential areas) can be also treated either as trip production (leaving home in the morning) or trip attractions (returning home in the afternoon). To determine transit ridership along a certain transit route, detailed land uses are examined in the Minneapolis-St. Paul metropolitan areas. Time-of-day count models are developed based on disaggregate data on parcel-level land uses and automatically collected farecard transactions, representing individuals’ movements. In this model, transit demand from farecard transaction data is considered as the dependent variable, and detailed land uses are used as explanatory variables. The model is able to explain the distribution of weekday boardings at a specific stop over time. This model can be used to manage and design effective transit services, such as increasing service along specific corridors.
  Authors: Lee, Sanggu; Hickman, Mark D.; Tong, Daoqin
  Authors: Lee, Sanggu; Hickman, Mark D.; Tong, Daoqin
  Year: 2013
  Document Type: Paper
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-4055
• Predicting Transit Ridership at Stop Level: Role of Service and Urban Form
  Abstract: This research aims to better understand the relative and combined influence of transit service characteristics and urban form on transit ridership at the stop level. Three metropolitan regions in Oregon were included in the analysis, representing different types of communities. We use stop-level ridership data from 7,214 TriMet stops in the Portland region, 1,400 Lane Transit District (LTD) stops in the Eugene-Springfield and 350 Rogue Valley Transit District (RVTD) stops in Jackson County (Medford-Ashland area) as the dependent variable for regression models. Categories of independent variables tested include: (1) socio-demographics; (2) transit service characteristics (e.g. headways, hours of service, transfer stops, bus vs. light rail, etc.); (3) land use (employment, population, land use type, pedestrian destinations, etc.); and (4) transportation system (e.g. street connectivity, bike lanes, etc.). The final model results indicate that the TriMet model does a better job explaining the variation in ridership at the stop-level; the adjusted-R2 is 0.69, compared to 0.61 for the LTD model, and 0.53 for the RVTD model. Land use characteristics around transit stops do have significant effects on transit ridership, though these effects are much smaller than the effects of transit level of service. Socio-demographic characteristics seem to have a larger effect on ridership in the large urban area than small urban areas (TriMet: 24% vs. LTD and RVTD: 11%). The land use characteristics have much smaller effect in large urban area than small urban area (TriMet: 5% vs. RVTD: 18%).
  Authors: Dill, Jennifer; Schlossberg, Marc A.; Ma, Liang; Meyer, Cody
  Authors: Dill, Jennifer; Schlossberg, Marc A.; Ma, Liang; Meyer, Cody
  Year: 2013
  Document Type: Paper
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-4693
  Practice-Ready: Yes
• The Time Between: Continuously Defined Accessibility Functions for Schedule-Based Transportation Systems
  Abstract: Accessibility is traditionally considered to be a property of a point or region in space, and to be invariant over time (or at least over some computationally convenient time interval). However, a location’s accessibility can vary over time on a wide range of scales. This temporal variation is especially significant for schedule-based transportation systems. Current integral measures of accessibility generally reflect the accessibility only at points in time corresponding to the departures of one or more trips; accessibility between these time points remains unconsidered and undefined. Consequently, these measures are insensitive to changes in route frequency and the distribution of trip departure times. Furthermore, these approaches ignore the disutility experienced by a system user who is limited to departing or arriving at scheduled times rather than at preferred times. As a result, they systematically overestimate the accessibility experienced by users of scheduled transportation systems. This paper establishes new methods for representing the accessibility provided by a schedule-based transportation system from a specific location as a continuously-defined accessibility function (CDAF) of desired departure time, defined for all time points. Using schedule and route information from metropolitan transit providers, applications of these methods are demonstrated to gain new insight into the accessibility provided by real-world transportation systems. Four examples are developed to represent common service types in metropolitan transit networks. The results confirm that accessibility is significantly overestimated by measuring single points and show that trip frequency is more valuable for sustained accessibility than high accessibility on individual trips.
  Authors: Anderson, Paul Richard; Owen, Andrew; Levinson, David M.
  Authors: Anderson, Paul Richard; Owen, Andrew; Levinson, David M.
  Year: 2013
  Document Type: Paper
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-4918
• Predicting Transit Ridership at Stop Level: Role of Service and Urban Form
  Authors: Ma, Liang
  Authors: Ma, Liang
  Year: 2013
  Document Type: Presentation; Poster
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-4693
• Stretching Resources: Sensitivity of Optimal Bus Frequency Allocation to Stop-Level Demand Elasticities
  Authors: Verbas, Ismail Omer
  Authors: Verbas, Ismail Omer
  Year: 2013
  Document Type: Presentation; Poster
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-5252
• Stretching Resources: Sensitivity of Optimal Bus Frequency Allocation to Stop-Level Demand Elasticities
  Abstract:

  Bus transit route frequencies in practice are often set reactively, without consideration of ridership elasticity to the service frequency provided. Where elasticities are used in frequency allocation, a single across the board value is typically used for all routes and all times of the day. The most advanced applications might use two values, for peak and off-peak respectively. With growing availability of ridership data from many sources, estimation of spatially and temporally disaggregated elasticities of demand with respect to service frequency. But do these make a difference in the resulting solution to the frequency allocation problem? This study is intended to examine this question by comparing the quality of solutions obtained using an optimal frequency allocation model with different sets of elastcities corresponding to varying levels of spatial and temporal disaggregation. Three main methodologies for estimating ridership elasticity with respect to headway are compared in the context of a Transit Network Frequency Setting framework: (1) temporal elasticities based on time of day, (2) spatio-temporal elasticities using a linear regression model and (3) categorical elasticities via grouping stops into demand, supply and land use classes. Elasticities based only on temporal aggregation may result in over-allocation of resources to certain time periods as compared to elasticities which account for some spatial characteristics, such as land use and the opportunity to transfer to other modes. It is also important to capture longer-term effects—over a year or more—in these models because seasonal activity patterns (e.g. school trips, vacation) may bias elasticity estimates over shorter time horizons. The experiments demonstrate that spatial detail in ridership elasticity estimation results in meaningful improvements in an objective function minimizing wait time and maximizing ridership, even when time periods are aggregated. Since much of this data is available at census tract level and collected by regional planning authorities, transit agencies could implement this frequency allocation formulation using rather coarse data.

  Authors: Verbas, Ismail Omer; Frei, Charlotte; Mahmassani, Hani S.; Chan, Raymond
  Authors: Verbas, Ismail Omer; Frei, Charlotte; Mahmassani, Hani S.; Chan, Raymond
  Year: 2013
  Document Type: Paper
  Subject: Planning and Forecasting; Public Transportation
  Session: 424
  Paper Number: 13-5252