61 Transportation Research Record: Journal of the Transportation Research Board, No. 2649, 2017, pp. 61–70. http://dx.doi.org/10.3141/2649-07 Monitoring speeds and identifying problem areas are essential for any public transport system because of the direct impact on its operating costs and on users’ travel time. This study generated a tool that iden- tified, quantified, and displayed operational bottlenecks of bus opera- tion in a city. The model was applied to the public transport system in Santiago, Chile, which faced a steady decline in operating speed. It was possible to identify locations with the most serious problems; this factor allowed transit authorities to focus their efforts on the areas that needed it the most. In addition, it was found that problems were concentrated in certain sectors of the city, including the central business district and intersections where the radial axis roads encountered the city’s central ring road. Once a problem is identified, it is essential to conduct site visits and combine the findings of this research with other sources of information to find the cause of the problem and propose solutions. One of the most important aspects of any bus system is operational speed, which is defined as the average speed of buses, considering dwell times at bus stops. The reasons for this importance are twofold: first, improvements in speed directly affect users’ travel time in terms of in-vehicle travel time savings plus possible waiting time savings and increased travel time reliability. Second, speed has a direct impact on the efficiency of the system: the same transport capacity can be achieved with a smaller fleet (reducing cost), or higher capacity can be achieved with the same fleet (implying increased frequency and therefore less waiting time). In order to improve speed, the first step is to monitor it. When large-scale automated vehicle location (AVL) data are used, most monitoring focuses on identifying streets with the lowest speeds. Often this monitoring is done at an aggregate level, which hinders researchers’ ability to properly characterize the problem (exactly where and when problems are arising) or to identify improvement thresholds (if the problem is solved, what speeds can be achieved). These thresholds are key to being able to identify projects with the highest return on investment, that is, those that will be the most beneficial to the system. The objective of this research is to develop a comprehensive tool to identify (a) where and when problems occur and (b) their respective improvement thresholds through the identification of bottlenecks. This study develops a simple model for identifying operational bottle- necks with data from ADATRAP (1–3). The model is then applied with real data from the public transit system, Transantiago, in Chile, identifying as a result the worst bottlenecks in the system. Then a tool to visualize bottlenecks in the city is presented, which allows easy identification of the spatial distribution of the problems and monitors the variation in magnitude between different periods of the day and the evolution in different years, among other applications. BACKGROUND This section consists of a brief review of literature in the field of speed monitoring and AVL data for public transport, with an empha- sis on the major causes of bottlenecks for bus operation. Then an overview of Santiago’s public transport system is given and the current operational context is described. Literature Review The significant uptick in the use of AVL data for public transport in recent decades (4, 5) has resulted in their wide use in major cities around the world. Turnbull defines different uses of this type of data in public transport systems (6). The most relevant uses are user information, planning, and monitoring and control of operations. Most studies on operations-monitoring for buses using GPS data have analyzed the reliability of bus travel times (4, 7, 8) and focused especially on compliance with itineraries. This type of analysis is relevant in cities with a high proportion of low-frequency service that operates on fixed bus stop arrival schedules. Okunieff presents the results of a survey of public transportation agencies, in which the biggest incentive for implementing AVL systems is to improve schedule adherence (4). The argument presented points to the direct improvements in user experience, information delivery, and service reliability. El-Geneidy et al. used archived intelligent Identifying and Visualizing Congestion Bottlenecks with Automated Vehicle Location Systems Application to Transantiago, Chile Christopher Bucknell, Alejandro Schmidt, Diego Cruz, and Juan Carlos Muñoz C. Bucknell, A. Schmidt, and J. C. Muñoz, Centro de Desarrollo Urbano Sustent- able, Departamento de Ingeniería de Transporte y Logística, Pontificia Universidad Católica de Chile, Avenida Vicuña Mackena 4860, Macul, Santiago 7820436, Chile. D. Cruz, Gerencia de Planificación y Desarrollo, Directorio de Transporte Público Metropolitano, Moneda 975, Santiago 8320239, Chile. Corresponding author: A. Schmidt, aschmid1@uc.cl.