Exploiting the Knowledge of Dynamics, Correlations and Causalities in the Performance of Different Road Paths for Enhancing Urban Transport Management

The great abundance of multi-sensor traffic data (traditional traffic data sources - loops, cameras and radars accompanied or even replaced by the most recent - Bluetooth detectors, GPS enabled floating car data) although offering the chance to exploit Big Data advantages in traffic planning, management and monitoring, has also opened the debate on data cleaning, fusion and interpretation techniques. The current paper concentrates on floating taxi data in the case of a Greek city, Thessaloniki city, and proposes the use of advanced spatiotemporal dynamics identification techniques among urban road paths for gaining a deep understanding of complex relations among them. The visualizations deriving from the advanced time series analysis proposed (hereinafter referred also as knowledge graphs) facilitate the understanding of the relations and the potential future reactions/outcomes of urban traffic management and calming interventions, enhances communication potentials (useful and consumable by any target group) and therefore add on the acceptability and effectiveness of decision making. The paper concludes in the proposal of an abstract Decision Support System to forecast, predict or potentially preempt any negative outcomes that could come from not looking directly to long datasets.