A Big Data Architecture for Traffic Forecasting Using Multi-Source Information

An important strand of predictive analytics for transport related applications is traffic forecasting. Accurate approximations of the state of transport networks in short, medium or long-term future horizons can be used for supporting traveller information, or traffic management systems. Traffic forecasting has been the focus of many researchers over the last two decades. Most of the existing works, focus on single point, corridor, or intersection based predictions with limited efforts to solutions that cover large metropolitan areas. In this work, an open big-data architecture for road traffic prediction in large metropolitan areas is proposed. The functional characteristics of the architecture, that allows processing of data from various sources, such as urban and inter-urban traffic data streams and social media, is investigated. Furthermore, its conceptual design using state-of-the-art computing technologies is realised.