|
TBEST Main Resource
|
|
TBEST stands for The Transit Boardings Estimation and Simulation Tool. TBEST has been interfaced with and derives its GIS functionality from ESRI’s ArcGIS suite of products. Over the past few years, the FDOT Public Transit Office has been at the forefront in the development of state-of-the-art transit analysis tools. In the specific area of transit ridership estimation and forecasting, the FDOT Public Transit Office funded the development of the Integrated Transit Demand and Supply Model (ITSUP) and the Regional Transit Feasibility Analysis and Simulation Tool (RTFAST). Both of these tools were well received by the user community for their ability to estimate ridership at the individual stop-level, user-friendly graphical interface, GIS functionality, and analytical rigor. The user community provided a wealth of feedback and helped identify a number of improvements that could be made to these tools. TBEST represents a third and culminating effort in the development of robust stop-level models of transit ridership. TBEST is a comprehensive transit analysis and ridership forecasting model that is capable of simulating travel demand at the individual stop-level while accounting for network connectivity, spatial and temporal accessibility, time-of-day variations, and route competition and complementarity.
TBEST simulates transit ridership at the level of the individual stop clearly distinguishing among stops at the same location, by route and direction. Thus, it is a “micro-level” model that can provide very detailed information regarding ridership estimates at individual stops. However, TBEST can also be used to obtain more aggregate route level, segment level, location-based, or system level measures through the aggregation of stop-level outputs. By simulating ridership at the level of the individual stop, the model intends to provide a strong platform and robust framework for modeling transit ridership in a region by time of day and day of week. The motivation behind the development of stop level forecasting was not necessarily a need for stop level forecasts, but rather the recognition that the predominant walk access nature of transit use requires a level of geographic precision that equates to understanding the size and characteristics of the activities and population in the stop level walk access buffer. Thus, TBEST has a model structure that is sensitive at the stop level even though service planning decisions are and should continue to be made at the route segment level as transit use behavior variations not explainable by stop level input variables are significant enough that users are discouraged from using individual stop level results for stop level service planning decisions. The current version of TBEST has been calibrated using year 2000 census and ridership data from Portland. Model estimation results and coefficients furnished in this users guide and provided as default parameters with the installation version of TBEST should be used with caution in other areas, particularly if the area is very different in character than Portland. While the model structure attempts to capture travel behavior as much as possible within the model variables as opposed to the constants, and the model is designed to calibrate the system total ridership to replicate local base conditions, caution is warranted, particularly as the experience base with TBEST is currently modest. Users who desire to use model coefficients customized to their own area will need to re-estimate and re-calibrate the various equations that comprise the TBEST model using census and stop-level ridership data from their own area. The modeling methodologies and estimation procedures are described in the Appendix of this users guide. In general, the user should note that the re-estimation and re-calibration of models comprising TBEST is a complex statistical and data mining exercise that requires considerable time and effort. This is not recommended unless there is extensive local TBEST use anticipated and applications experience has not proven satisfactory with default coefficients. |
