Evaluating Travel Time Accuracy in Navigation Systems
Aditya Parekh
Co-Presenters: Individual Presentation
College: The Dorothy and George Hennings College of Science, Mathematics and Technology
Major: Computer Science
Faculty Research Mentor: Daehan Kwak
Abstract:
GPS-based navigation applications, such as Google Maps, Waze, BingMaps, and MapQuest, have become part of our daily travel, offering users real-time traffic information and ETAs to help construct the best routes possible. Yet, the accuracy of such ETAs in relation to observed ground truth data is under-explored, raising questions about the reliability of these systems in making route decisions. This study aims to determine the quality of traffic information provided by major navigation services by comparing ETAs with observed travel times from physical sensors and traffic data systems.The methodology used in this work has two major components. First, we design a Python-based web mining system to collect real-time traffic data, including estimated time of arrivals (ETAs) from four well-known navigation platforms: Google Maps, Waze, MapQuest, and BingMaps. Second, we collect ground truth traffic data from publicly available sources, including electronic toll tag readers and loop detectors provided by the New York Department of Transportation (NY DOT). The datasets are classified based on specific time frames, which include rush hours, non-rush hours, and weekends.We conduct a case study of densely populated urban areas, using New York City as the experimental setting. Using statistical methods, we compare differences between predicted and actual travel times. Descriptive statistics are also used to identify trends and anomalies across different navigation systems.Our results aim to improve the knowledge of ETA reliability, helping users make better-informed route decisions, while also enabling navigation providers to enhance their systems. These insights can be integrated into a user-friendly dashboard, bridging the gap between raw traffic data and actionable route recommendations. This work contributes to advancing technology in intelligent transportation systems and interdisciplinary research in data science, computer science, and smart cities.