Traffic speed estimation based on multimodal data fusion and graph analytics

Public bus transportation plays a significant part in Singaporeans’ daily commuting lives. Thus, it is important to ensure that the service quality (e.g. bus travel time and frequency) of bus trips are maintained at the best level to meet our commuters’ demand. One major factor that affects the service quality is weather condition. Wet weather condition (e.g. heavy rains) would significantly slow down the bus travel speed and hence increase the bus travel time. Longer commuting time for our passengers could lead to dissatisfaction in our public bus service if not properly handled. In this project, we aim to investigate the effect of rainfall on bus travel speed using historical CEPAS data. We will also build a machine learning model to predict future bus travel time by taking training features such as number of passengers, time of the day and most importantly, weather condition into account. Our prediction results are essential to forecast the bus travel time accurately, so that we can implement solutions when the predicted travel time is slower than usual. By anticipating the longer commuting time in advance and devising solutions such as scheduling more buses to solve the issue as soon as predicted, commuter’s travelling time will not be affected heavily during wet weather. Therefore, this ensures that the public bus transportation service quality in Singapore will not be compromised in the event of heavy rain. In short, our project’s objective is to maintain Singapore’s public bus transportation service quality during wet weather, to provide a delightful and beyond satisfactory journey for all our commuters.Bachelor of Engineering (Computer Science