Data Analytics for Dependable Transportation Systems in a Smart City

Public transit is an important component of the day-to-day activities of many people. It provides a cost-effective and convenient way for individuals to commute to work, school, and other destinations. Bus transit is a vital mode of transportation for students, as it enables them to commute to and from their educational institutions. Delays in bus schedules can have severe consequences-such as missing exams, meetings, and other important engagements-in daily activities of city residents. Hence, in this paper, we present a data science solution for mining and transportation analytics on public transit on-time performance data. Knowledge discovered from these data helps improve public transit performance, and thus enhance rider experience in a city. This helps build a smart city. To elaborate, our solution adapts frequent pattern mining, which identify uncover variations in transit performance across various neighborhoods. Through identifying significant findings, we establish correlations to determine the factors contributing to bus delays in specific areas. Improving the bus arrival or departure time can have a positive impact on the overall usability and attractiveness of bus transit for commuters since people are more likely to use buses when they can rely on them to arrive on time and get to their destinations promptly. Our solution also provides users features to visualize the discovered knowledge about the bus departure time in all the neighborhoods at different times of the day. Evaluation results on real-life public transit data from a Canadian city demonstrated the practicality of our data science solution towards the building of a dependable transportation system in a smart city.