Experience with Data Quality in the Transport Domain

The chapter describes the design and application of a methodology for evaluating the quality of traffic data. It emphasizes its importance for the entire transport sector and presents specific procedures and tools that make it possible to detect low-quality data. This increases user trust and improves the effectiveness of traffic management. The methodology is based on the concept of data governance. The introduction describes the current state, initial conditions, and the general ten-step process for data quality assessment. The evaluation itself is divided into syntactic and semantic checks. This includes statistical methods for detecting anomalies in the data, allowing the identification of errors or unusual traffic situations. The main part of the chapter focuses on the application of the methodology to traffic data. It provides a detailed description of the design, implementation, and use of different quality rules, along with their logical aggregation. This enables hierarchical evaluation of data quality, from individual traffic lanes to the entire road network, both at a detailed and comprehensive level of quality assessment. The functionality of the methodology and all partial procedures is illustrated by applying it and all proposed quality rules to real data from a selected section of the D11 highway in the Czech Republic in 2023. The results of the evaluation and their interpretation are presented. The types of anomalies detected in traffic volume and average speed data are also discussed. It demonstrates that automated rule-based systems can significantly improve the reliability and usability of traffic data.