Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
77 Effectively capturing, archiving, and using AVL-APC data involves overcoming organizational as well as technological challenges. This chapter summarizes the chief challenges reported by the surveyed agencies. 12.1 Raising the Profile of Archived Data A reason that many AVL systems have failed to deliver their potential in terms of useful archived data is that those who specified and designed the systems either did not emphasize the importance of archived data or, more likely, did not recognize important differences between the needs of real-time data and those of archived data (38). Time and again, procurements have focused on real-time applications, with the implicit expec- tation that archived data analysis would somehow happen. Some of the lack of appreciation of the character and value of archived data has been on the part of vendors whose primary product is real-time information; and some has been on the part of transit agency staff who managed procurements. There is a need for transit agency staff who are involved in system pro- curement to better understand how system design affects what data is captured, what the data quality will be, and what off-line analyses it will be able to support. At the same time, there is a need for decision makers to appreciate the importance of archived data acquisition and analysis for improving system management and performance. Several studies have looked for quantifiable benefits of AVL systems to justify their cost. Where benefits have been quan- tified, they most often come from an off-line application, namely, revising scheduled running times. This is ironic, con- sidering that off-line analysis has often been an afterthought of such systems. 12.2 Management Practices to Support Data Quality Control and supervision has traditionally been concerned about performance, not about data collection. In fact, to the extent that buses are ordered to deviate from their schedules, operations control makes schedule matching more difficult. Fortunately, agencies are becoming more aware that col- lecting and later analyzing data can also contribute to improv- ing performance. There is an opportunity to improve data quality by changing control and supervision practices. Exam- ples include detecting and correcting invalid sign-on data, informing the AVL-APC system of a revised schedule when a bus is deviated from its schedule, and standardizing codes for control messages. 12.3 Staffing and Skill Needs To date, transit agencies making good use of AVL-APC data have been able to do so only because of the strong set of staff skills they have been able to employ. The lack of available data analysis tools has meant that agencies have needed the skills to develop their own database and analysis tools. Only because of dedicated, qualified, and resourceful staff were Tri-Met and King County Metro able to make the great strides they did in improving the quality of their AVL-APC data and in develop- ing tools to analyze it. In the future, with the development of third-party analysis software and increasingly relevant and accepted standards for data definitions and interfaces, the need for expertise in infor- mation technology (IT) to develop archived AVL-APC data systems should decline. However, agencies will still need the staff and expertise to analyze the data. Managing data quality takes considerable staff effort. At agencies with good AVL and APC data, one or two staff mem- bers are usually devoted to overseeing that the systems deliver the data they should. Identifying and correcting accuracy problems sometimes takes considerable IT expertise, espe- cially if matching algorithms or data objects have to be changed. With AVL systems, matching, accuracy, and data capture issues are often just as much a problem for real-time applica- tions as for archived data, so that little additional work is needed for archived data itself. C H A P T E R 1 2 Organizational Issues
12.4 Managing an Instrumented Sub-fleet Where APCs are used, common practice is to equip only 10% to 15% of the fleet. The logistical issues of managing an equipped sub-fleet are significant. A program manager is needed to make assignments of instrumented buses and to check whether assignments were made. Garage supervisors must stage the instrumented buses properly, and transporta- tion supervisors must ensure they are used where assigned. Data collection efforts that require all the buses operating on a route or along a trunk be instrumented at the same time (e.g., for a headway analysis) are particularly demanding; if a few instrumented buses miss their assignments, the data col- lection effort will have to be repeated another day. Another layer of complication is added if so many instrumented buses are needed in one place that intergarage transfers are needed. Aging of the equipped sub-fleet also poses problems. At first, the instrumented buses are new, and there may be polit- ical demands to assign the new buses in ways that conflict with a data collection program. As they age, restrictions on using them on certain routes or runs can develop. For exam- ple, they may not have the ergonomic seats needed by some operators, or the low floor required on some routes. They may develop maintenance problems with wheelchair lifts or other bus systems. Partly to avoid the complications of shifting instrumented buses around, Tri-Met is well on its way to equipping its entire fleet with APCs. With 65% of the fleet instrumented with APCs (and the entire fleet instrumented with AVL), Tri-Met simply allows the APC-instrumented buses to collect data wherever they get assigned; that gives it an adequate sample size. When APCs can be integrated into a smart bus system that already includes AVL and schedule matching, the mar- ginal cost of adding APCs drops dramatically. 12.5 Avoiding Labor Opposition Suspicion of âthe spy in the cabâ or âbig brotherâ is natu- ral. There is always the danger of sabotage if transit operators resent the way a system monitors them, especially if they believe that the system is unfair or inaccurate. For the most part, transit agencies that have adopted AVL and APCs have avoided incapacitating labor opposition. Gen- erally, agencies have been successful at communicating the security benefits of AVL, which builds operator support. APCs generally do not engender opposition, perhaps because their name suggests they are only counting passengers. Some agencies intentionally avoid directly challenging an operator with AVL-APC data. However, some use the data to identify patterns of abuse or poor performance, alerting supervisors where and when problems are likely to occur. 78
