Developing a Data Quality Evaluation Framework for Sewer Inspection Data

• Published in: Water (Basel) Vol. 15; no. 11; p. 2043
• Main Authors: Khaleghian, Hossein, Shan, Yongwei
• Format: Journal Article
• Language: English
• Published: Basel MDPI AG 27.05.2023
• Subjects: Algorithms; Artificial intelligence; Asset management; assets; Data analysis; Data collection; data quality; Data science; Decision making; Information management; Infrastructure; Inspection; Inspections; Machine learning; Quality control equipment; Quality management; Sewer systems; water
• DOI: 10.3390/w15112043
• ISSN: 2073-4441

The increasing amount of data and the growing use of them in the information era have raised questions about the quality of data and its impact on the decision-making process. Currently, the importance of high-quality data is widely recognized by researchers and decision-makers. Sewer inspection data have been collected for over three decades, but the reliability of the data was questionable. It was estimated that between 25% and 50% of sewer inspection data is not usable due to data quality problems. In order to address reliability problems, a data quality evaluation framework is developed. Data quality evaluation is a multi-dimensional concept that includes both subjective perceptions and objective measurements. Five data quality metrics were defined to assess different quality dimensions of the sewer inspection data, including Accuracy, Consistency, Completeness, Uniqueness, and Validity. These data quality metrics were calculated for the collected sewer inspection data, and it was found that consistency and uniqueness are the major problems based on the current practices with sewer pipeline inspection. This paper contributes to the overall body of knowledge by providing a robust data quality evaluation framework for sewer system data for the first time, which will result in quality data for sewer asset management.