CONSTRUCTION BUSINESS & PROJECT MANAGEMENT CONFERENCE · Vol. 2021

Requirements for the application of data sciences in construction

JA Wium and CJ Jurgens

Keywords: Data sciences, project efficiency, rework, project parameters

Abstract

This paper investigates the use of data sciences in construction project management. A literature review identified limited use and application of data sciences in construction project management compared to other industries. Therefore, the paper addresses this gap by demonstrating the benefits of using data mining as a project management tool. It identifies the requirement for structured data capturing, facilitating the widespread application of data sciences in construction with demonstrated benefits. Following a description of the data mining methodology used, two cases are presented of the practical application of data mining on construction projects. In one case, predictions are made for the number of employment opportunities created in a project. In the other case, the total project duration is estimated/predicted. It is shown that improved results (in comparison with previously published cases) can be obtained from large databases with a 62.3% accuracy in one case study and 88.8% accuracy in the other. It also shows that the accuracy depends on the chosen algorithm. Similar to previous studies, the two cases identified the effort required to prepare suitable databases. This is a result of the current unstructured and uncoordinated capturing of data in the construction industry. It is recommended that companies develop standardised processes to capture data. It will be even more beneficial if such data can be stored in an industry-wide database to improve management processes and predictions. A currentlimitation to its application is the lack of knowledge in the industry about data sciences techniques.